THE ROLE OF SOCIAL CAPITAL IN THE INTERNATIONALISATION PROCESS OF EGYPTIAN AND UK SMES: A CO- EVOLUTIONARY PERSPECTIVE

Transcription

1 THE ROLE OF SOCIAL CAPITAL IN THE INTERNATIONALISATION PROCESS OF EGYPTIAN AND UK SMES: A CO- EVOLUTIONARY PERSPECTIVE By ROSE NAROOZ A thesis submitted to The University of Birmingham For the degree of DOCTOR OF PHILOSOPHY Department of Marketing Business School University of Birmingham January 2014

2 University of Birmingham Research Archive e-theses repository This unpublished thesis/dissertation is copyright of the author and/or third parties. The intellectual property rights of the author or third parties in respect of this work are as defined by The Copyright Designs and Patents Act 1988 or as modified by any successor legislation. Any use made of information contained in this thesis/dissertation must be in accordance with that legislation and must be properly acknowledged. Further distribution or reproduction in any format is prohibited without the permission of the copyright holder.

3 Abstract There is increasing evidence of the positive role that the social capital offered by attachment to networks plays in facilitating the internationalisation of SMEs. Network social capital serves as a key platform for small firms to access resources and competencies they lack. Yet surprisingly little is known about how network social capital co-evolves with the international growth and survival of SMEs. In other words, what happens after the initial move into internationalisation and how do SME networks develop over time? The present study aims to address this gap through providing a comprehensive analysis of internationalisation patterns, network social capital development and how these two fields may co-evolve over time. Moreover, little is known about how SMEs from different institutional contexts rely on their network attachments to cope with the challenges and opportunities presented by the external environment. The study addresses this gap and compares well-established SMEs from two national contexts (UK and Egypt) which contrast in their cultures and levels of institutional development. This study relies on comparative multiple case studies consisting of twenty UK and twenty Egyptian SMEs. These case studies investigate different internationalisation and networking patterns and identify co-evolutionary patterns of internationalisation and associated networking behaviour over time. A mixed methods approach to data analysis was used to investigate co-evolutionary patterns and inter-country differences respectively. The study contributes to the International Entrepreneurship [IE] literature through offering a comprehensive and dynamic understanding of internationalisation and by indicating key differences in the behaviour of SMEs from emerging and developed economies. It concludes that the co-evolutionary perspective usefully highlights the dynamic interaction between ii

4 SMEs and their contexts. It shows that SMEs not only adapt to changes in their external environment but in some cases are able to influence the environment to their favour. The thesis closes with implications and recommendations for Egyptian and UK policy makers and managers. iii

5 To Hoda, Ibrahim and Joy iv

6 Acknowledgements "The whole of science is nothing more than a refinement of everyday thinking" -Albert Einstein I am truly indebted and thankful for my kind, supportive and inspiring supervisor. Professor John Child is more than just a supervisor he is a second father to me. Without his constructive feedback, guidance and direction this PhD would have not come to light. Professor Child is no doubt one of the best academics in the field and also he is one of a great heart and kindness. He sincerely cared for my wellbeing and professional growth. I am entirely thankful for his support and thank God each day for having such a supervisor. I also owe special thanks for Dr. Svetla Marinova and Professor Suzanna Rodrigues who helped me during the first years of my PhD. Both have been very supportive both on academic and professional levels. I felt I had two older sisters who understood the tough times a PhD student could pass through. This dissertation would not have been possible without the help, sacrifice and support of my parents and sister. They have offered support, help, and sacrifice in every possible loving ways. I owe everything to them. I am fortune to have such a supportive and loving family. Also, I owe sincere and earnest thankfulness to Mr. and Mrs. Mountford have taken me under their wing and treated me more than a daughter. They have been my father and mother here in the UK. I sincerely thank God for having such a caring and loving people in my life. v

7 I owe special thanks and gratitude to my amazing friends and colleagues. Nora, my best friend who have always been there by my side in all the good and bad times, Daniella, who has the most kind and loving heart, she is a second sister to me. Mayada, who helped and encouraged me especially during my final submission. Shaker, I owe you special thanks for all of your brotherly help and support. I also owe special thanks for Vima, Sima, Paulo, Sammy and all of my colleagues. It brings tears to my eyes to be surrounded by such loving and supportive colleagues and friends. vi

8 Table of Contents Chapter 1: Introduction and overview 1 Introduction 1 Theoretical background 4 Research questions 7 Research approach and empirical investigation 8 Structure of the thesis 10 Chapter 2: Literature review 12 Introduction 12 Definitions of SMEs internationalisation Internationalisation models and perspectives 19 Uppsala internationalisation model (U-model) 23 Innovation related models (I-model) 27 Network model of internationalisation 28 International Entrepreneurship as an emerging field International new ventures: a challenge to the traditional stage and process models of internationalisation Oviatt and McDougall's (1994) new venture typology 38 Other SME typologies 44 Summary of internationalization models and perspectives Theories of social capital 48 Social capital definitions and perspectives 50 vii

9 Benefits and costs of social capital 55 Networks and social capital 57 Ego networks 58 Evolution of social capital 62 Research on networks and the internationalisation of SMEs Historical evolution of institutional theory 69 Institutions: definitions and Scott's institutional pillars 72 Institutions, social capital and the entrepreneurial behaviour of firms 74 National institutional differences between the UK and Egypt 77 Institutional theory and its dimensions 78 Contextual differences, SME internationalisation and networking 81 Cognitive-cultural informal institutions 81 Normative informal institutions 94 Formal institutions The co-evolutionary perspective 100 Co-evolution of firms and external environment 101 Co-evolution and SMEs internationalisation and networking behaviours Summary and Conclusion Chapter 3: Proposed theoretical framework 114 Introduction 114 New venture co-evolutionary model 117 The pre-internationalisation stage [PIS] 117 Initial network social capital 118 viii

10 Contextual effects 119 Early internationalisation stage [EIS] 120 The subsequent internationalisation stage [SIS] 121 Structural capital evolution 122 Structural capital evolution at the overall network level 122 Evolution of dyadic relationships 123 Relational social capital development 128 The traditional SME co-evolutionary model 129 Start-up and domestic expansion stages 130 International expansion stage 130 Conclusion 133 Chapter 4: Methodology 135 Introduction 135 Theoretical orientation 135 Epistemological and ontological orientations 137 Research strategy 145 Validity, reliability and replicability 147 Case study design 153 Defining network boundaries and sampling 161 Research design and methods used in this study 166 Data arrangement and preparation 170 Data analysis 175 Summary of the research process 177 ix

11 Conclusion 180 Chapter 5: Overall co-evolutionary patterns: networks and internationalisation development patterns 181 Introduction 181 Generic framework and data preparation 182 A. UK SMEs evolutionary and co-evolutionary patterns 193 A.1 Within stage correlations and overall development patterns 193 A.2 Across stages co-evolutionary correlations 213 B. Egyptian SMEs evolutionary and co-evolutionary patterns 218 B.1 Within stage correlations and overall development patterns 219 B.2 Across stages co-evolutionary correlations 239 Speed of internationalisation as an alternative perspective of internationalisation process 244 Discussion and conclusion 257 Chapter 6: SMEs taxonomies and clusters 262 Introduction 262 Generic networking and internationalisation differences between UK and Egyptian SMEs 264 A. Proposed SMEs clustering 267 A.1 Clustering of UK SMEs 270 A.2 Clustering of Egyptian SMEs 272 B. Taxonomies definitions and operationalisation of measures 275 B.1 UK SMEs taxonomies 280 I) Limited internationaliser/networkers 280 x

12 II) Extensive internationalisers/networkers 287 III) Moderate internationalisers/networkers 295 B.2 Egyptian SMEs taxonomies 307 I) Extensive internationalisers/networkers 308 II) Moderate internationalisers/networkers 318 III) Hybrid internationalisers/networkers 326 Conclusion 334 Chapter 7: Antecedents of SMEs internationalisation and networking process 335 Introduction 335 Home country institutional context 336 Egyptian institutional characteristics 340 UK institutional characteristics 357 Entrepreneur's international experience 365 Conclusion 369 Chapter 8: Discussion and conclusion 371 Introduction SME internationalisation patterns SME network development patterns SME internationalisation and co-evolutionary patterns Antecedents and contingencies of SME internationalisation and networking behaviour 393 Revised research framework 398 Managerial implications 404 xi

13 Implications for policy makers 406 Limitations and directions for future research 408 Appendices 413 Bibliography 456 xii

14 List of Tables Table 2.1: Internationalisation definitions 15 Table 2.2 Internationalisation models and perspectives 20 Table 2.3: Johanson and Mattsson (1988) typology 30 Table 2.4 Key themes examining the importance of networks in SME internationalisation 66 Table 2.5: The development of institutional theory 70 Table 2.6 Relevance of institutional development for internationalization and networking 75 Table 2.7: Indicators of the extent of institutional voids in Egypt and the UK 80 Table 2.8: Examples of government programs and support offered to UK SMEs 97 Table 2.9: GEM entrepreneurial framework conditions (EFC) 99 Table 4.1: Shortcomings of quantitative research 146 Table 4.2: Shortcomings of qualitative research 147 Table 4.3: Types of validity 148 Table 4.4: Alternative criteria for evaluating qualitative research 150 Table 4.5: Rationales for Using Case Studies 156 Table 4.6: Advantages and limitations of case studies 157 Table 4.7: Comparison between different sampling approaches 162 Table 4.8: Sources of UK and Egyptian SMEs 166 Table 4.9: Individual tie coding based on different tie characteristics-the Manu5 case. 172 Table 4.10: Triangulation of research data. 176 Table 5.1: Definitions and operationalisation of key variables 186 Table 5.2: Average network measures within stages 194 xiii

15 Table 5.3: Within stage network correlations (Spearman's rho correlation coefficient) 202 Table 5.4 Internationalisation outcomes within stages 204 Table 5.5: Within stage internationalisation correlations (Spearman's rho correlation coefficient) 209 Table 5.6: Within stage co-evolutionary correlations (Spearman's rho correlation coefficient) 211 Table 5.7: Co-evolutionary correlations across stages (Spearman's rho correlation coefficient) 215 Table 5.8: Average/mean values of network variables within stages 219 Table 5.9: Within stage network correlations (Spearman's rho correlation coefficient) 221 Table 5.10: Average/mean values of internationalization outcomes within stages 227 Table 5.11: Within stage internationalisation correlations: 229 Table 5.12: Within stage co-evolutionary correlations (Spearman's rho correlation coefficient) 232 Table 5.13: Across stages Co-evolutionary measures (Spearman's rho correlation coefficient) 240 Table 5.14: Details of classification of SMEs at different stages of internationalization 247 Table 5.15: SMEs descriptive 247 Table 5.16: Internationalisation outcomes per stage of early and traditional SME 248 Table 5.17 Table 5.18: Table 5.19 Table 5.20 Correlation between speed of internationalization and internationalization outcomes of UK SMEs (Spearman's rho correlation coefficient) Correlation between speed of internationalization and internationalization outcomes of Egyptian SMEs (Spearman's rho correlation coefficient) Correlation between speed of internationalization and network variables of UK SMEs (Spearman's rho correlation coefficient) Correlation between speed of internationalization and network variables of Egyptian SMEs (Spearman's rho correlation coefficient) Table 6.1: Operational definitions 277 Table 6.2: Tie characteristics and operational definitions 279 xiv

16 Table 6.3: Table 6.4: Table 6.5: Table 6.6: Table 6.7: Table 6.8: Table 6.9: Internationalisation and networking processes of limited internationaliser/networker. Internationalisation and networking processes of extensive internatinalisers/networker Internationalisation and networking processes of moderate internationaliser/networker Internationalization and networking processes of extensive internationaliser/networker Internationalisation and networking processes of extensive internationaliser/networker Internationalisation and networking processes of moderate internationaliser/networker Internationalisation and networking processes of hybrid internationaliser/networker Table 7.1: Key institutional/contextual characteristics 341 Table 7.2: Role of Egyptian supporting agencies 353 Table 7.3: Key institutional characteristics in UK 361 Table 7.4: The role of UK supporting agencies 362 Table 7.5: Assessment of UK and Egyptian entrepreneurs prior international experience 366 Table 7.6: Cross tabulation of the presence of UK entrepreneurs' prior international experience and the type of support used. 367 Table 7.7: Cross tabulation of the presence of Egyptian entrepreneurs prior international experience and the type of support used Table 7.8: Cross tabulation for the complied sample of UK and Egyptian SMEs 368 Table 7.9: Chi -square test for the complied UK and Egyptian SMEs 368 Table 8.1: Recommendations for newly internationalising SMEs xv

17 List of Figures Figure 1.1: Structure of thesis 11 Figure 2.1: Casual Cycle adapted from Johanson and Vahlne (1977). 25 Figure 2.2: Types of new ventures 42 Figure 2.3: An example of ego net (adapted from Arenius, 2002) 59 Figure 2.4: Scores for Egypt and the UK on Hofstede s Cultural Dimensions 92 Figure 2.5: The co-evolutionary perspective as a framework to study the development of SME internationalisation and networking over time 109 Figure 3.1: New venture co-evolutionary model 115 Figure 3.2: Traditional SME co-evolutionary model 116 Figure 4.1: Burrell and Morgan s (1979) assumptions underlying nature of social sciences 139 Figure 4.2: Technical definition of case-study as an all-encompassing method 154 Figure 4.3: A traditional quantitative method and its limitations 161 Figure 4.4: Illustrative example of using mixed methods to enhance triangulation 165 Figure 4.5: SME s internationalisation stages 174 Figure 5.1: Generic framework 183 Figure 5.2: Within network development patterns 195 Figure 5.3: Within stages internationalisation development patterns 207 Figure 5.4: Within stage network development patterns 220 Figure 5.5: Within stage internationalisation development patterns 228 Figure 5.6: UK SMEs internationalisation patterns based on speed of early and subsequent internationalisation 250 xvi

18 Figure 5.7: Egyptian SMEs internationalisation patterns based on speed of early and subsequent internationalisation 251 Figure 5.8: UK SMEs co-evolutionary model 258 Figure 5.9: Egyptian SMEs co-evolutionary model 259 Figure 6.1: UK SMEs dendrogram. 270 Figure 6.2 UK SMEs taxonomy 272 Figure 6.3: Egyptian SMEs dendrogram. 273 Figure 6.4: Egyptian SMEs taxonomy 274 Figure 6.5: Figure 6.6: Figure 6.7: Figure 6.8: Figure 6.9: A Summary of internationalisation profiles and network attachment of UK limited internationaliser/networker taxonomy A Summary of internationalisation profiles and network attachment of UK extensive internationaliser/networker taxonomy A Summary of internationalisation profiles and network attachments of UK moderate internationaliser/networker taxonomy A summary of the internationalisation profiles and network attachments of Egyptian extensive internationaliser/networker taxonomy A summary of internationalisation profiles and network attachments of Egyptian moderate internationaliser/networker taxonomy Figure 6.10: A summary of internationalisation profiles and network attachments of Egyptian hybrid internationaliser/networker 333 Figure 8.1: Internationalisation co-evolution of New Ventures 399 Figure 8.2: Internationalisation co-evolution of Traditional SMEs 400 Figure 8.3: Revised analytical framework 401 xvii

19 List of Abbreviations PIS EIS SIS SC PFS No. of Markets No. of Regions Av.Net Density Av.Net Constraint Av. Net Size MNC(s) SME(s) IMTs SCTs CTs Pre-Internationalisation Stage Early Internationalisation stage Subsequent Internationalisation Stage Social Capital Percentage of Foreign Sales Number of Markets Number of Regions Average Network Density Average Network Constraint Average Network Size Multinational Corporation(s) Small and medium-sized Enterprise(s) Intentionally managed ties Semi-coincidental ties Coincidental ties xviii

20 CHAPTER 1 INTRODUCTION AND OVERVIEW Introduction This research is concerned with examining the impact of network social capital 1 on the internationalisation process of small and medium-sized enterprises [SMEs]. I became interested in this research topic after coming across an Egyptian entrepreneur who had no previous international experience and limited resources but who managed within the course of a few years to export into more than ten markets. Like most Egyptian entrepreneurs he was faced with challenges and obstacles that made it nearly impossible to operate either domestically or internationally. When I asked him how he managed to overcome such obstacles and internationalise in such a short time, he attributed the company s success to his personal relationships. These relationships helped him to acquire the necessary capital to start up his business. Moreover, he had friends who were scattered across different middle-eastern countries and through them he managed to sell his products. At the time I started this research before the 25 th January 2011 revolution it was rare to find SMEs that were internationally successful and even more unusual to find SMEs that continued to survive after their initial start. So the question of how SMEs in my country, Egypt, can facilitate their internationalisation through networking came to interest me. 1 Network social capital refers to the resources that are made readily available through an actor s network of relationships - family/friends, suppliers, buyers, competitors, organisations, research institutes, export promotion agencies and colleagues. In other words, these are the benefits that an actor can access through his/her network of relationships whether s/he is directly involved, semi-involved in the creation of these relationships or the tie creation as a result of a coincidence. These ties could vary in strength, type and duration 1

21 I began looking at the literature of SMEs internationalisation and found a plethora of studies on SMEs internationalisation and the importance of network social capital in facilitating their internationalisation. After reviewing different strands of literature, such as industrial networks, social network, embeddeness, social capital, internationalisation and institutional theories, it was clear that this is a mature field. Also, it is a well-established fact that networks and the social capital they create are crucial for the SMEs internationalisation. However, I also realized that almost all the research evidence gathered so far relates to developed western countries rather than to developing ones like Egypt. This raised the question of how the experience of Egyptian SMEs would compare with that of the ones from a western country such as the UK. I also began to see that there were key gaps in the literature. For example, most studies had focused on the early years of the SME life cycle. Several authors were calling for more research into what happens after the initial exporting or internationalisation stage (Jones, Coviello and Tang, 2011; Kuivalainen, Sundqvist, Saarenketo and McNaughton, 2012). Thus, studies investigating the impact of networks on SME internationalisation had focused on the early stages of SMEs life cycles; relatively few had investigated subsequent stages, such as growth and maturity stages (Coviello, 2006 ;Hoang and Antoncic, 2003; Slotte-Kock and Coviello, 2010). These gaps in the literature pointed to the need for a dynamic perspective on how SME networking and internationalisation processes worked out over time. I was puzzled as to which one affects the other. Is an SME s network social capital a key determinant of its internationalisation at different stages of its life cycle? Or is it, rather, the SME s internationalisation choices which determine how its networking develops over time? 2

22 This thesis is informed by the view that these two processes internationalisation and networking change over time and that they are better understood through a dynamic lens. In other words, that they co-evolve together, influencing one another at different points in time, and that they are affected by the wider external environment, such as the conditions pertaining to the country where an SME is located. Entrepreneurs, as I mentioned in the opening example, have to find ways to cope with these conditions and survive. Based on my limited knowledge at the early stages of this research, the only literature in which I could find some attention to this issue was that on organisational adaptation to the environment The organisational adaptation literature suggests that firms adapt to the environment(s) in which they are embedded through adopting two approaches. The first is a reactive approach whereby they respond to changes in the environment as these emerge. The second approach is a proactive approach whereby companies anticipate which changes are likely to take place and plan accordingly. I struggled with this line of reasoning, especially how I could apply it to the case of SMEs. I then came across Rodrigues and Child s (2009) book on corporate coevolution, which uses co-evolution as a perspective to explain how the firms adapt and even change external environment through leveraging their internal processes, resources and capabilities in such a way that enables them to have superior performance compared to their competitors (see also Volberda and Lewin, 2003). I also discovered that the co-evolutionary perspective has been used by other scholars in evolutionary economics and organisational theory (see Child, 2012 for an overview). Co-evolution has its roots in biology and it is, strictly speaking, concerned with the way two organisms co-develop and influence one another at different points in time (Anderson and May, 1982). These organisms influence and are influenced by their host environment. This 3

23 provides a perspective that crosses different levels of analysis and takes into account the dynamic nature of the processes in a firm s development. In other words, the co-evolution perspective can potentially serve as a guiding framework for investigating the relationship over time of an SME s internationalisation and networking and whether both of these evolving patterns are themselves shaped by the institutional environment in which an SME is embedded. Theoretical background Recent studies of SME internationalisation focussed particularly on rapidly internationalising firms that seek to achieve competitive advantage through operating in multiple overseas markets immediately after their inception or soon afterwards (Oviatt and McDougall, 1994; Oviatt and McDougall, 2000; Rugman and Verbeke, 2007; Chetty and Agndal, 2007; Young, Dimitratos and Dana, 2003). These SMEs are often referred to as born globals, new ventures and early start ups (Oviatt and McDougall, 1994:2005; Prashantham and Young, 2011; Coviello, 2006). Other patterns of SME internationalisation have also been identified. For example, Bell et al. (2003) found out that certain SMEs the so-called born again global focus on expanding domestically and only after a certain time then begin to rapidly internationalise into overseas markets. These types of SMEs constitute challenge to the traditional internationalisation theories applied to multinational corporations [MNCs], whereby the internationalisation process is seen to be gradual and sequential (Johanson and Vahlne, 1977; Oviatt and McDougall, 2005a; 2005b; Madsen and Servias, 1997). The main premise of sequential models of internationalisation of which the best known is the stage or Uppsala model is that firms go through a set of rationally-based behavioural steps. They start off with a minimal resource commitment and then increase their involvement as their 4

24 level of market knowledge increases (Johanson and Wiedersheim-Paul, 1975; Johanson and Vahlne, 1977; 1990). Luostarinen (1970; 1979) found similar results, arguing that the firm increases its degree of international commitment as its level of experiential knowledge increases. Compared to MNCs, SMEs have limited resources and capabilities which in principle makes it difficult for them to engage in international operations. SMEs are said to suffer from a liability of smallness (Knight and Cavusgil, 1996; Rugman and Verbeke, 2007; Chetty and Agndal, 2007; Coviello and Murno, 1997). In addition, SMEs tend to have limited international experience and therefore suffer from a liability of newness (Aldrich and Auster, 1986; Han, 2006). This means that the success of these firms depends to a great extent on their ability to acquire relevant resources from other organizations and internalise those competencies in which they excel. Oviatt and McDougall (1994) argue that SMEs rely on other governance mechanisms such as strategic alliances and networks to access valuable resources without participating in their creation. These networks of relationships serve as a key platform through which SME could access resources and knowledge they need for their internationalisation (Coviello and Murno, 1997; Bell, McNaughton and Young, 2003; Jones, Coviello and Tang, 2011; Kuivalainen, Sundqvist, Saarenketo and McNaughton, 2012). Child and Rodrigues (2011) argue that smaller firms are particularly likely to rely on the "mediation" of others in coping with external complexity and uncertainty. In other words, network relationships create social capital for the SMEs, which is a resource that fosters ease of movement and access to resources as well as helping them to spot opportunities as they emerge in foreign markets (Arenius, 2002; Coviello, 2006; Sharma and Blomstermo, 2003). 5

25 Accordingly, firm-level social capital is vital for the survival and growth of SMEs. Most of the literature to date, however, has focused on early internationalisation, providing insights on how networks of social capital influence entry mode, market selection and speed of internationalisation. In addition, they offer some understanding of how social capital evolves within the first years of SME operation. However, less is known about "how and why the network develops in a certain way" (Slotte-Kock and Coviello, 2010:51). Furthermore, empirical studies to date have tended to rely mainly on cross-sectional designs and to a lesser extent on case studies. They have treated networks either as a dependent or an independent variable, whereby they are seen either as a key facilitator for the SMEs internationalisation or as being shaped by the internationalisation choices of SMEs. They have not considered that the two may co-evolve. Moreover, the entrepreneurial process is more meaningful when examined within its broader context, namely the institutional environment as this could assist (or alternatively hinder) internationalising SMEs. Emerging economies typically have weak institutions and tend to suffer from institutional voids. These voids result from absent or poorly-performing specialized intermediaries (such as export promotion agencies), regulatory systems and contract-enforcing mechanisms that govern day-to-day transactions (Khanna and Palepu, 1997; Khanna and Palepu, 2010). Even if formal institutional arrangements are in place, there remain significant "gaps between formal rules and norms, and their enforcement in daily practice" (Rodrigues, 2013: 14). So far, scholars have paid little attention to the implication of these domestic institutional voids on the internationalisation of SMEs and how entrepreneurs react to fill in these voids through relying on formal and informal means of support. 6

26 In particular in countries where relationships are the way of doing business especially in collectivist cultures entrepreneurs are more likely to rely on informal means (their personal relationships) to fill in institutional voids inherent in their domestic market. By contrast, entrepreneurs from countries that place less value on relationships and focus more on formal means of governance and support are more likely to depend on contractual based relationships. This will not only impact the SMEs domestic operations but it will have a bearing on how SMEs in general create networks of relationships in overseas markets. Accordingly, institutional differences are expected to influence SMEs networking and internationalisation behaviours. Based on my identification of key gaps in the current literature, I sought a guiding framework that would allow me the flexibility to use different theories and disciplines to explain phenomenon under investigation, and at different levels of analysis. In particular, I needed a framework that would allow for investigation of the dynamic nature of SMEs internationalisation and networking processes over time and to investigate how they codevelop. The co-evolutionary perspective seemed to meet this requirement. Research Questions The general objective of this dissertation is to contribute to the international entrepreneurship literature, which is at the interface between the internationalisation literature and entrepreneurship literature. It draws from theories that touch on this interface, such as internationalisation theory, social capital theory and institutional theory. In the main, studies 7

27 have treated social capital like other forms of capital, viewing it as an asset that a firm possesses and manages in a way that allows it to extract value (Leana and Van Buren, 1999). Accordingly, the key research questions I formulated in the context of a comparison between SMEs in Egypt and the UK are: (1) What are the different internationalisation patterns of Egyptian and UK SMEs over time? (2) What are the network development patterns of Egyptian and UK SMEs over time? (3) Is there co-evolution between the internationalisation and network development patterns of Egyptian and UK SMEs over time? If so, what are the underlying reasons for the co-evolution between the internationalisation and network development patterns based on the understanding and interpretations provided by the UK and Egyptian SME decision makers? (4) What are the key antecedents or contingencies of SMEs internationalisation and networking behaviours? Research approach and empirical investigation As stated in the four research aims, the study presented in this thesis aims to understand the dynamics of the SMEs behaviour through investigating how network social capital facilitates their internationalisation process and how the internationalisation choices made by SME decision-makers may shape the firm s network development over time. An additional interest is how key contingencies such as the domestic institutional context influence the network attachments and internationalisation profiles of SMEs. 8

28 This research draws on multiple case studies of Egyptian and UK SMEs. The companies selected for study are ones that have been in operation for at least 10 years, have less than 250 employees, and have been operating in overseas markets for a period of at least 6 years. Since the focus of this research is on subsequent years after the initial first internationalisation attempt, I focused on well-established SMEs. The study included 40 case studies 20 from the UK and Egypt respectively and including both manufacturing and service firms. The main method of data collection was semi-structured interviews which were conducted with founder(s) and mid-level professionals who had first-hand experience with the international operations of the firm. There has been a tendency in the international entrepreneurship literature to focus on hightech companies (such as software and bio-tech firms), which usually specialize on a narrow market niche and follow an accelerated internationalisation pattern. I wanted to incorporate SMEs from different sectors both high and low-tech firms to allow sufficient variation to appear in their networking and internationalisation profiles so as to be able to identify different evolutionary and co-evolutionary patterns. The research is designed to capture the subjective perceptions of the entrepreneurs and adopts a retrospective view about the internationalisation history of the firm, the creation and evolution of social capital and sequence of events. It uses a mixed method approach. Firstly, I use non-parametric statistical tests to identify key patterns of association between internationalisation and networking development. 2 Secondly, I focus on rich qualitative evidence to understand the underlying mechanisms of internationalisation and networking processes through developing novel exploratory clustering co-evolutionary taxonomies of 2 This was done though focusing on three structural dimensions of network and three internationalisation performance variables within and across difference stages of SME s lifecycle. 9

29 SMEs that help to produce a deeper understanding of each of the two processes and how they co-develop together. The use of qualitative and quantitative methodologies offered complementary perspectives on the SME s behaviour over time. The scope of my inquiry also permitted an analysis across different levels national context, firm and individual decision-maker. Structure of the thesis The structure of thesis is shown in Figure 1.1. It divides into five main parts. The first part contains this introduction and overview. The second part, containing Chapters 2 and 3, covers theoretical background and a priori analytical framework. It also sets the scene through providing the rationale behind the choice of UK versus Egypt as research contexts. The third part discusses the key research approach and methods used (Chapter 4).The fourth part presents the key empirical findings in Chapters 5, 6 and 7. Finally, the fifth part (Chapter 8) covers the discussion of key findings, contributions and implications of research. 10

30 Figure 1.1: Structure of thesis Empirical Findings Chapter 1 Introduction and overview -Introduction and background -Research questions -Research approach and empirical investigation -Thesis structure Theoretical Background Chapter 2 Literature review -Internationalisation theories/models and typologies of SMEs -Social capital definitions vs networks, embeddness and industrial networks definitions -Social capital development -Institutional context and national differences between UK and Egyptian contexts -Co-evolutionary perspective Chapter 3: Theoretical framework -Development of research questions -Development of a priori frameworks -Traditional SMEs coevolutionary model - New venture s coevolutionary model. Chapter 4 Research methodology -Philosophical orientation -Research approach and design -Data collection -Data preparation and arrangement -Data analysis approach Chapter 5 Overall Coevolutionary patterns -Data analysis generic framework -Operationalisation of measures -UK SMEs evolutionary and co-evolutionary patterns (within &across stages) -Egyptian SMEs evolutionary and co-evolutionary patterns (within & across stages) -Speed of internationalisation as an alternative perspective of SMEs internationalisation process. Chapter 6 SMEs Taxonomies and Clusters -Generic differences between UK and Egyptian SMEs -Proposed exploratory clustering of SMES -Clustering of UK & Egyptian SMEs -Taxonomies definitions and operationalisation of measures -UK SMEs taxonomy -Egyptian SMEs taxonomy Chapter 7: Antecedents of SMEs internationalisation and networking processes -Internationalisation and networking antecedents Home country institutional context -Egyptian & UK institutional context -Entrepreneur s previous international experience - Link between network attachments, presence of entrepreneurs previous international experience and internalisation profiles of SMEs Chapter 8 : Discussion and conclusions -Key findings in relation to the four research questions -Contributions to the existing body of knowledge -Methodological contributions -Comparison between apriori and revised frameworks -Implications for managers and policymakers -Research limitations and suggestions for future research 11

31 CHAPTER 2 LITERATURE REVIEW Introduction The central role of SMEs in driving economic development at national and international levels has been widely recognised. SMEs growth serves as an engine for promoting innovation and enhancing the economic development of countries (Reynold, 1997; Busentiz, Gomez and Spencer, 2000). SMEs grow through expanding into overseas markets. Governments and policy makers pay significant attention to entrepreneurs and small business owners. They attempt to assist and support small companies in their domestic and international growth. There is an extensive literature on SME internationalisation that can be traced back to the 1970s. However, most of these studies have focused on a snap shot of SME internationalisation, focusing on the first exporting attempts into overseas markets (Welch and Paavilainen-Mäntymäki, 2014; McAuley, 2010). Internationalisation is not just about initial export or entry into foreign markets, rather it is a dynamic process through which an SME increases its international commitment and operations and the process by which it moves into foreign markets and adapts its operations to suit overseas markets (Welch and Luostarinen, 1993; 1998; Calof and Beamish, 1995; Johanson and Mattsson, 1993). Recent reviews of SMEs internationalisation have called for the time dimension to be incorporated and for understanding internationalisation as a process through which SMEs engage in 12

32 overseas operations (c.f. Welch and Paavilainen-Mäntymäki, 2014; Kuivalainen, Sundqvist, Saarenketo and McNaughton 2012a; Kuivalainen, Sundqvist, Saarenketo and McNaughton, 2013; Ruzzier, Hisrich and Antoncic, 2006; Harris and Li, 2005; Jones, Coviello and Tang, 2011). Given that SMEs have limited resources and international experience, they tend to rely on network social capital to the obtain resources and support they need to engage in overseas operations (Sasi and Arenius 2008; Coviello, 2006; Slotte-Kock and Coviello, 2010). Network social capital refers to resources that are made readily available through an actor s network of relationships such as, family/friends, suppliers, buyers, competitors, organisations, research institutes, export promotion agencies and colleagues. In other words, these are the benefits that an actor can access through his/her network of relationships whether s/he is directly involved, semi-involved in the creation of these relationships or the tie creation as a result of a coincidence. These ties could vary in strength, type and duration 3. This highlights that importance of networks and the social capital they create for the SMEs' international growth. Furthermore, studies that investigated the impact of networks on SMEs internationalisation focused on the early stages of an SME s life cycle, with only a few taking account of the subsequent stages, such as the growth and maturity stages (Coviello,2006, Hoang and Antoncic, 2003; Slotte-Kock and Coviello, 2010). These major gaps in the literature indicate the need for a dynamic view of SMEs networking and internationalisation processes. 3 From the above definition there are two dimensions of network social capital. These are the structural (presence of ties) and the relational dimensions. Each type of network relationship varies in strength and benefits offered to focal actor. These benefits change and/or diminish over time. Moreover, this definition covers different types of relationships with varying degrees strength and different networking logics. This is different from Huggins (2010) definition of the network social capital. He defines network social as a rational decision to invest in relationships (mainly inter-firm relationships) that enables the firm to gain access to knowledge to enhance their "expected economic gains" (p.336). Moreover, he differentiates between network social capital and social capital. The latter involves investment in social relationships through which individuals gain access to social returns (such as emotional support). Whereas the former, involves rational decision taken by the actor to invest in calculative networks to achieve economic gains and meet business-related targets. 13

33 SMEs do not operate in a vacuum; they are influenced by the domestic and international environments in which they operate. SMEs do not necessarily increase the level of their international commitment in a linear manner. They could shrink their international operations (de-internationalise) and re-internationalise at a later stage (Kuivalainen, Saarenketo, Puumalainen, 2012a; Kuivalainen et al., 2012b). SMEs are highly influenced by the broader context in which they operate and they are subject to pressures from foreign governments, and their industry and domestic institutional context (Kuivalainen et al., 2013; Dodd and Patra, 2002; Pajunen and Maunula 2008; Kiss and Danis, 2008). These external pressures are also likely to impact on how the SME uses and draws on its network social capital to access resources and on the support it needs to adapt to these external pressures (Welch and Paavilainen-Mäntymäki, 2014). This calls for a framework that allows for a "contextualised view" of the SME s internationalisation and networking dynamics and a framework that crosses different levels of analysis (Welch and Paavilainen-Mäntymäki, 2014:15; Pajunen and Maunula, 2008). This chapter focuses on addressing these aspects through reviewing different streams of literature or theories used to inform the present research. These are primarily internationalisation, networking /social capital, institutional and co-evolutionary theories. Firstly, I summarise the key definitions and models of internationalisation and the internationalisation of different types of SMEs. Secondly, I consider different theories pertaining to networks and social capital and compare these theories with concepts such as embeddeness and industrial network. I discuss the concept of social capital in detail and the different dimensions proposed by Nahapiet and Ghoshal (1998). Thirdly, I present a brief description of institutional theory, especially its origins and key dimensions. This informs a discussion of the key contextual differences between the two countries involved in this study 14

34 (Egypt and the UK) and the rationale behind choosing these two countries as the main research context. Fourthly, I summarize the origins and background of the co-evolutionary perspective, which will be the guiding framework for this study. 2.1 Definitions of SME internationalisation Existing research on SME internationalisation has adopted several definitions. For instance, Nordic models of internationalisation traditionally defined internationalisation as "a process of increasing involvement in international operations" (Welch and Luostarinen, 1993:156) which is usually regarded as an incremental and gradual process. A firm can increase its involvement in international operations (Luostarinen 1979; Johanson and Wiedersheim-Paul, 1975) but could also at a particular point of time reduce its "international involvement and de-internationalise" (Ruzzier et al, 2006: 478; Calof and Beamish, 1995). Other scholars have focused on a firm s international operations through analyzing its products, mode of entry and markets (Luostrarien, 1979; Ruzzier et al, 2006), and applied other perspectives, such as the network perspective (Johanson and Mattsson, 1988; 1993). Some of the definitions used by different scholars are identified in Table 2.1. Table 2.1: Internationalisation definitions Key authors Johanson Valhne (1990) Johanson Mattsson (1993) and and Internationalisation definition A cumulative process, in which relationships are continuously established, developed, maintained and dissolved in order to achieve the firm s objectives. The process of adapting the firms operations (in terms of strategy, structure and resources) to international environments. Welch and Luosteinen (1993) Calof and Beamish (1995) The outward movement of firm s international operations. The process of increasing involvement in international operations. 15

35 Ahokangas (1998) Lethinen and Penttinen (1999) The process of mobilising, accumulating and developing resources stocks for international activities. It is concerned with developing networks of business relationships in other countries through extension, penetration and integration. *Source: Ruzzier et al., It is also concerned with the relationship between the firm and its internal environment. It derives its origin from the development and utilisation process of different international activities, primarily inward, outward and cooperative operations. Johanson and Mattsson (1988) defined internationalisation as "a cumulative process, whereby by relationships are continually developed, established, maintained and broken and dissolved in order to achieve the objectives of the firm" (p.306). A major shortcoming of this definition is that it focused exclusively on relationships. Johanson and Mattsson (1993) subsequently refined their definition arguing that internationalisation process involves adapting a firm s internal resources and strategies to international environments. Lehtinen and Penttinen (1999) attempted to gather most of the important features of the internationalisation process into one definition based on the results of the Nordic models. They incorporated two key dimensions of internationalisation namely international orientation and commitment (Ruzzier et al, 2006). The former refers to the perceived psychic distance between overseas markets and home market (Reid, 1981) while the latter emphasises the amount of resource commitment that a particular mode of operation requires. It is evident that some scholars have shifted their attention from international activities and modes of operation towards the resources needed for internationalisation, following the 16

36 resource based view of the firm. For instance, Ahokangas (1998) argues that internationalisation involves movement, accumulation and creation of resources needed to carry out international activities. He proposes that the SME s internationalisation depends on the natural context, for example networks, in which the SME is embedded that provide it with the required resources, and he focuses to a lesser extent on the international activities of the firm (Ruzzier et al., 2006). A holistic view of the international involvement of the firm would include selling products to overseas markets, importing products from abroad and even a joint or collaborative agreement between the firm and a foreign partner (Korhonen, 1999). This is reflected in Lehtinen and Penttinen s (1999) definition. Others have focused implicitly or explicitly on the dynamic nature of SMEs internationalisation and argued that it is a multi-dimensional process which evolves over time. This process involves a number of internationalisation decisions and events (Jones and Coviello, 2005; Welch and Paavilainen-Mäntymäki, 2014). The understanding of SME internationalisation behaviour has been widely examined in the literature through focusing on the scale or scope of internationalisation. A major shortcoming of these studies is that they focus on internationalisation at a single point of time, or on a single dimension of internationalisation (Sullivan, 1994). For example, some studies have focused on the degree of international intensity/scale, which refers to percentage of foreign sales that comes from foreign operations (Preece, Miles and Baetz, 1999). Others have focused on the extent of internationalisation, which refers to number of value-adding activities that are carried outside in foreign markets (Bloodgood, Sapienza and Almeida, 1996); or on the scope of internationalisation, which refers to the number of 17

37 countries to which the firm exports (Zahra, Ireland and Hitt, 2000; McNaughton, 2003). One study that began to focus on the multi-dimensionality of internationalisation is Fisher and Reuber (1997). They argue that the degree of internationalisation is composed of: (1) percentage of foreign sales to total sales. (2) Percentage of firm s employees who spend more than 50% of their time on international activities, and (3) geographical scope of the SME s activities. The main premise of my research is that internationalisation is a dynamic process that takes place over time. Therefore, I will define internationalisation as the process through which the firm increases its level of international operations - inward, outward and cooperative- over time. Moreover, I shall measure an SME's international performance through using Fisher and Reuber s (1997) dimensions which are: internationalisation scale (the percentage of foreign sales to total sales) in case of outward and cooperative strategies and internationalisation scope, which is the number of countries in which the firm operates. I add a third dimension, that is the number of regions in which the SME operates, as this could be an indicator of the degree of international diversity or concentration of SME s overseas operations (Kuivalainen, Sundqvist and Servais, 2007; Kuivalainen et al., 2012a; 2012b). So far I have discussed the internationalisation definitions used by different scholars. Different models of internationalisation, their weaknesses and theoretical underpinnings are discussed in the following section. 18

38 Internationalisation models and perspectives Several perspectives and models have been used by scholars. The first group of studies are often referred to as process theories of internationalisation (Andersen, 1993). The most commonly cited models are the Uppsala model (U-model) and the Innovation-related model (I-Model). The second stream of studies is concerned with a network approach to internationalisation which was initiated by the industrial marketing and purchasing group (IMP) researchers (Johnson and Mattsson, 1988; 1985). The third set of studies draw upon resource and dynamic based views of the firm, and have been used to explain the internationalisation process of different types of SMEs (mainly traditional and new ventures).the final group of studies are those on international new ventures or born global companies, which began to emerge in the early 1990s (such as Oviatt and McDougall, 1994; 1995; 1997; Bloodgood et al., 1996; Fischer and Reuber, 1997; Autio, Sapienza, Almeida, 2000; Knight and Cavusgil, 2005). Each of these perspectives is presented in Table 2.2 and will be discussed in the following section. 19

39 Table 2.2. Internationalisation models and perspectives Uppsala and Nordic/learning Models RBV/dynamic capabilities Network model Innovation-related models New Venture Theoretical roots Behavioural theory of firm (Cyert & March, 1963; Ahokangas, 1998); and Penrose's theory of knowledge and change in organizations (1959). Rooted in strategic management (Andrews, 1971) and Penrose s theory of the growth of the firm (1959). Rooted in Resource dependency theory and extension of the Uppsala model (Johanson and Mattsson, 1988). Derived from the work of (Rogers, 1962; Gankema, Snuif and Zwart, 2000). Entrepreneurship and resource based view (Jones et al, 2011, Zahra and George, 2002) Main premise Markets are in equilibrium and managers tend to avoid uncertainty. Focuses on sustainable and unique imitable resources that create sustainable competitive advantage for the firm The firm s ability to defend its market position through attaining and keeping its relevant resources (such as knowledge based resources) and on dynamic ability to learn and develop new resources. Tries to avoid uncertainly. Recognizes that the firm is dependent on resources owned by other firms in the industrial network. Each subsequent stage of internationalisation is an innovation for a firm that focuses exclusively on exporting. Has fixed stages ranging from 3-6.(broadly pre-export, initial export and advanced exporting stages) Markets are imperfect and the entrepreneurs unique set of resources (knowledge, relationships, experience and ability to coordinate resources) are the main reasons these firms succeed Entrepreneurs are alert for opportunities Level analysis of Firm and environment Firm resources and capabilities Firm-Firm (micro level or Firm to network (macro level) Individual and firm Entrepreneur (mainly). Approach: reactive/proac tive Reactive Proactive Reactive Reactive Proactive Triggers of internationalisation. Experiential knowledge and learning. International expansion Knowledge development, resources and market position Learning Entrepreneurial vision 20

40 Speed Slow Depends Slow-fast (typologies) Slow Fast Entry mode Following the stages or establishment chain model. For example, firms start with direct exporting and/or agents and as their level of international experience increases they will use more advanced modes of entry such as joint ventures or wholly owned subsidiaries in overseas markets N/A. Stages ranging from direct to indirect contingent on network (production and external networks) Stages Mixed entry modes. Distance markets to Low psychic distance Mostly low distance markets May go for distant markets if nearby ones offer no opportunity. Low psychic distance Mainly distant markets (companies form emerging or developed markets) Role network of N/A If combined with a network perspective, networks serve as means of securing access to resources. Network position is a crucial asset that could be leveraged N/A Entrepreneurs can modify and choose the networks they need to access needed resources -Firms could also engage in joint development strategies of imitable unique resources. Criticism Deterministic (Reid, 1981) Individuals have no strategic choice (Andersson,2000) Challenged by emergence of NVs The resources must be transparent (Grant, 1991); durable and so forth. The nature of the resources tends to be vague and broadly defined with no clear boundaries between them. Firms use wider network of relationships to internationalise rather than just a production network. This could be with customers, suppliers, export promotion agencies, friends, consultants and competitors (Chetty and Holm, 2000). Theoretically imprecise The timing of different stages is not clear Determining the stage difference in terms of activities involved is Still a new field. Considered mainly with reference to international entrepreneurship only. 21

41 Ignores the firm decision-maker and firm characteristics when pursuing internationalisation opportunities Ignores the liability side (downside) of networks. subjective. Only describes the process of change rather than the dimensions on how firms develop their activities within each stage. *Sources: (Antoncic and Hisrich, 2000; Arenius 2002; Ruzzier et al., 2006) 22

42 Uppsala internationalisation model (U-model) This model first emerged in the work of Johanson and Vahlne (1977), and Johanson and Wiedersheim-Paul (1975). It has its theoretical roots in the work of Penrose's (1959) growth theory of the firm and in the behavioural theory of the firm (Aharoni, 1966). The main premise of the Uppsala model is that managers are risk averse people who tend to avoid uncertainty as much as possible. In other words, since managers lack foreign market knowledge, they perceive the risk of operating in foreign markets to be high. They therefore increase their level of international commitment in a gradual incremental manner as their level of experiential learning increase. This contrasts with the eclectic paradigm, where managers are expected to have a good level of knowledge and to be rational individuals who make rational decisions regarding the ownership and location advantages (Dunning, 1988; 2000). Another key premise of the Uppsala model is that markets differ in terms of their psychic distance, which refers to the degree to which the culture, language, business norms and economic development of countries are different (Johanson and Wiedersheim-Paul, 1975). It is generally assumed that the more distant a market is geographically the more its level of psychic distance increases. Accordingly, it becomes difficult for the firms to understand the markets and they have imperfect knowledge about the market and the level of uncertainty is high (Arenius, 2002). Nowadays, the institutional dimension of psychic distance appears to be more significant (Child, Rodrigues and Frynas, 2009). Johanson and Wisdersheim-Paul (1975) proposed that firms usually start to enter markets that have a minimal level of psychic distance and then incrementally move into more distant markets as their level of experiential learning increases. Moreover, the firm s involvement in 23

43 foreign markets takes the form of an establishment chain (Johanson and Vahlne, 1977; Johanson and Wisdersheim-Paul, 1975). This means that firms start off with an entry mode that requires the minimum amount of resource commitment such as exporting. Over time, as the level of their experiential knowledge increases, they become more confident to enter new markets. This means that they can move to distant markets with an entry mode that has a greater degree of resource commitment such as establishing a green-field manufacturing facility. Recent studies have lent some support to this model with respect to small firms. Examples are Crick and Jones (2000) and Holmlund and Kock (1998). The former found that firms follow incremental steps in internationalisation because managers tend to be risk averse and increase their level of commitment only after they gain the necessary international experience. The latter study posited that the most important modes of entry are through sales agents and direct exporting. This may indicate that small firms usually rely on sales representatives or agents as the primary mode of entry, since this requires the minimum amount of resources and carries the lowest level of perceived risk. On the other hand, Bell (1995) posited that a firm s internationalisation process is "less deterministic than stage models imply" (Arenius, 2002: 28). In a similar vein Jones (1999) found little support for the conventional stages presented by the model. Moreover, the Uppsala stage model follows a "causal cycle" where market knowledge and level of resource commitment in foreign markets has an impact on which mode of entry to use (commitment decision) in foreign markets and how they are going to perform their activities (Johanson and Vahlne, 1990:11). This is illustrated in Figure

44 Market knowledge Commitment Decision Market commitment Current Activities Figure 2.1: Casual Cycle adapted from Johanson and Vahlne (1977). Market knowledge and market commitment represent the firm s current state. They verify and alter the company s commitment decisions in foreign markets (Johnson and Vahlne, 1977). Johanson and Vahlne (1990) argued that Uppsala model applies to most companies with three main exceptions. The first is large companies which have excess resources enabling them to take bigger internationalisation steps; they tend to jump several steps in one go. This implies that the Uppsala model is more applicable to smaller sized firms. The second exception is if the conditions in the market are stable and knowledge can be easily obtained by similar companies. In this instance, the firm is more likely to obtain this knowledge through means other than through first-hand experience. Finally, if the firm operates in similar markets it can reuse its experience in those similar markets; it will have secured transferable knowledge. Weaknesses of Uppsala model Several authors have criticized process theories of internationalisation, particularly the Uppsala model (Madsen and Servais, 1997; Leonidou and Katsikeas, 1996). These scholars have questioned several theories underpinning its viability. Firstly, Madsen and Servais (1997) argued that most of the founders of new ventures tend to have prior market knowledge. This permits them to take on additional market commitments that are not specifically slow and gradual. Thus they argue that new ventures tend to follow a path of accelerated internationalisation. In a similar vein, Leonidou and Katsikeas (1996) argued that 25

45 the momentum through which the firm shifts from one stage to another and within each stage has not been taken into account by most of the stage models. Secondly, these models tend to ignore the strategic choices of individual managers. An organisation s position and place within a particular context is usually a result of choices made by its managers (Child, 1972). In addition, Child and Rodrigues (2011) argued that firms could alter the environment in which they operate though exerting power and influence over actors (with whom they have a relationship) in a way that creates benefits for them. This contradicts the deterministic view of the Uppsala model (Arenius, 2002). Thirdly, other forms of learning that may be available for firms are ignored (Forsgren, 2002). This could be achieved through engaging in exploring new alternatives and ideas, copying a best practice, and so forth. The Uppsala model tends to adopt a reactive form of learning, whereby it is only through first-hand experience with the market and process of trial and error that the firm increases its learning. This allows for other patterns of internationalisation to emerge that are not necessarily slow incremental ones (Forsgren, 2002). Finally the emergence of NVs and born global companies has cast considerable doubt on the applicability of these models. McDougall et al. (1994) presented such firms as ones that seek a competitive advantage through internationalising directly after their establishment or soon afterwards. They internationalise rapidly rather than slowly and tend to enter distant markets from the beginning. 26

46 Innovation-related Models (I-model) These models have their roots in the behavioural theory of the firm (Ruzzier et al., 2006). I- Models tend also to highlight the fact that firms follow a set of incremental steps because of their lack of experiential knowledge about markets and the high degree of uncertainty inherent in internationalisation decision (Andersen, 1993). The I-model is derived from innovation theories (Rogers, 1962; Robertson, 1971). These argue that innovation consists of sequential learning steps, whereby firms start off by adopting a new product or service. They then move to the next step through accepting or refusing an innovation as a result of the decision-makers cognitive and behavioural processes. When applying these models to firms, the first entry to export markets is a form of innovation by itself, such as developing a new product or idea. Thus entering foreign markets is like developing a new product and so forth (Andersen, 1993). For instance Reid (1981) has divided the innovation stages of an internationalising firm into 5 stages: export awareness, export intention, export trial, export evaluation and acceptance or rejection of exporting. A key weakness of the innovation model is that there are a large number of factors that enable the firm to move from one stage to another and influence the export behaviour of the firm (Bilkey, 1978). Both the Uppsala model and the innovation models apply to large and small firms and they both argue that firms tend to follow slow incremental steps when it internationalises. Thus they are similar in their theoretical underpinnings and suffer from similar weaknesses. 27

47 Network model of internationalisation The network model focuses on industrial markets and focuses on inter-organisational relationships as the key unit of analysis. The industrial market is viewed as a network of relationships between the firm, suppliers and customers and the model emphasises some dependencies and development processes that take place in foreign markets (Johanson and Mattsson, 1988, Arenius, 2002). This model is deeply rooted in resource-dependency theory. It emphasizes the fact that firms are dependent on other firms which have resources they need. This means that for a transaction to take place, firms need to establish relationships with other firms to access the resources they need. Firms are highly dependent on one another as they need to coordinate their activities in some way which is usually done through repeated interaction rather than a formal structure (Arenius, 2002). This is crucial for the firm s survival and growth, whereby firms engage in cross border activities and coordinate their resources in order to secure their survival. Johanson and Mattsson s (1988) network model is based on some key assumptions. Firstly, within industrial networks companies are linked to one another through a web of long term relationships and that previous exchange has taken place between them. Secondly, the network ties range from knowledge, legal, technical, economic and social. Thirdly, domestic relationships tend to be more developed and well established than foreign ones. Finally, that relationships change through repeated interaction and exchanges (Arenius, 2002). Thus, as will be discussed later, the firm s position within the production network and in different markets depends on its existing relationships with other firms. Thus relationships are path dependent and highly contingent on the firm s history. This has a bearing on the firm s future activities and plans. 28

48 Mattsson and Johanson (1988) have focused on two levels of analysis: the micro and macro levels. The former refers to the individual firm s position in the network, the volume of transaction/sales (relative importance of the firm) and the strength of relationships it has with its partners (Johanson and Mattsson, 1988). The macro level on the other hand, focuses on the role of the firm in the whole network, the weight of the network (such as market share) and firm identity. A network position is said to be advantageous for the firm if it offers access to resources and knowledge, but the process of securing this position is time consuming and requires a lot of investment in building and maintaining relationships. When it becomes well developed and formed it is viewed as "a market asset" (Johansson and Mattsson, 1988). The assumption is that as firms internationalise they create and maintain relationships with actors/partners in other countries. Over time the strength and the number of relationships between members of the network increases. This is achieved through one or more of the following strategies: 1. Market extension: the firm forms relationships with partners in countries that are new to it. 2. Penetration strategy: firms increase their level of commitment with already established foreign networks. 3. International integration: firms combine and mix their positions in the networks in various countries. The key underlying assumption is that firms need resources that their counterparts have and that they rely on network relationships to access these resources and to access foreign markets (Chetty and Holm, 2000). They classified the networks into two main types: production nets and national nets. The production network involves the firm s relationships 29

49 with its counterparts in related business areas, or with firms conducting similar or complementary activities. The national network, on the other hand, involves the firm s networks in other countries. Accordingly, Johanson and Mattsson (1988) classified firms into four main categories along two dimensions: the degree of internationalisation of the firm and the degree of internationalisation of the market. These are the early starter, the lonely international, the late starter and the international among others. The four different positions these firms could occupy are presented in Table 2.3. Table 2.3: Johanson and Mattsson s (1988) typology Degree of internationalisation of the market Low High Degree of internationalisation of the firm Low The early starter The late starter High *Adapted from Johanson and Mattsson (1988) The lonely international The international among others 1. Early starter: The firm, its competitors and suppliers have little knowledge about foreign markets and cannot rely on their domestic relationships to gain this knowledge. SMEs rely on their foreign relationships (agents and international clients) to enter foreign markets through providing them with the required market knowledge. This reduces the costs and risks associated with uncertainty. The introduction to 30

50 foreign markets is usually done through clients or distributors in foreign markets (Chetty and Holm, 2000). 2. Lonely international: is a company that is highly internationalised (has operations in several markets) but its market/industry still has a domestic focus. Johanson and Mattsson (1988) referred to it as a lonely international because it is the only firm (as opposed to its counterpart) that has the capabilities to internationalise. This firm or entrepreneur is assumed to have prior international experience in dealing with foreign markets, hence the "the Lonely international firm has an edge over its domestic competitors, as it has already established a position in the business network" (Chetty and Holm, 2000: 81). 3. Late starter: The market or industry is already internationalised and the firm is indirectly connected to foreign markets through its suppliers, competitors and clients (who have international operations in foreign markets). It is assumed that late starter firms find it difficult to enter close markets with a lower degree for psychic distance (as they are dominated by its domestic competitors) and that they therefore try to target more distant markets. They suffer from late mover disadvantages, because it is difficult to break into existing networks while high entry barriers are present. 4. International among others: The firm is highly internationalised and operates in a highly internationalised market. These firms have gathered and acquired the necessary international experience over time, this makes it easier to set up sales subsidiaries in foreign markets and increases the firm s ability to control and coordinate its operations across markets. These firms are highly involved with 31

51 networks across markets, which provide the firm with resources and information that are crucial for its foreign survival and growth. The main premise of the conventional network model of internationalisation is to understand the context in which entrepreneurial firm operates, which includes market conditions, and the firm s network of relationships, national and production networks (Madsen & Servais, 1997). Weaknesses of the Johanson and Mattsson (1988) model Chetty and Holm (2000) have identified several weaknesses when operationalising the Johanson/Mattson model to SMEs in New Zealand. Firstly, the criteria used to differentiate between each type of firm in the matrix tend to overlap. For instance, the early starter usually depends on introductions by clients or suppliers to enter other foreign markets. This could happen to other types of firms such as those which are already international among others. Secondly, relationships are not always an advantage because there is a liability side to relationships, as found by Coviello and Murno (1997) whereby relationships with some partners may inhibit and hinder the internationalisation of SMEs to other markets. Thirdly, the model does not clarify how firms shift positions in the matrix. How does an early starter, for example, shift to international among others? What is the process through which a firm moves from one part of the matrix to another? What is the role of relationships and how do they facilitate or hinder the transition from one position to another? Finally, the model ignores the relationships that are outside the production net such as governmental bodies that 32

52 offer support to SMEs (through export promotion programs) and customers who could serve as catalyst for internationalisation. International entrepreneurship as an emerging field The stage model is based on the economic perspective of internationalisation, particularly transaction cost theory. It ignores the process aspect of internationalisation (Vahlne and Nordstrom, 1993; Ruzzier et al., 2006). The process approach is useful in explaining key factors inherent in internationalisation but fails to acknowledge the individual strategic choices undertaken by entrepreneurs and managers of firms (Reid, 1981, Anderson, 2000). Entrepreneurs play a vital role in internationalisation of SMEs (Ruzzier et al., 2006). They tend to think in an entrepreneurial and strategic manner and focus on creating a fit between the firm s strengths and weaknesses with the external opportunities and threats (Hitt et al., 2001) 4. This has led to the emergence of International Entrepreneurship (IE) as a separate stream covering a different approach to SME internationalisation and it represents an intersection between the International Business and Entrepreneurship literatures (McDougall and Oviatt, 2000a; Antoncic and Hisrich, 2000). McDougall and Oviatt (2000) defined entrepreneurship as "combination of innovative, risk-seeking behaviour that crosses national borders and is intended to create value in organizations" (Ruzzier et al, 2006: 489). Although its definition is still vague and its application has been varied, some scholars extend this definition to refer to younger ventures -and to a lesser extent established SMEs (Acs et al., 2001; Ruzzier et al, 2006). In general, entrepreneurship can exist in large and well established firms as well as within small firms as it is an important mechanism for any firm s development (Ruzzier et al, 2006). 4 This is similar to the SWOT analysis in the strategic management literature. 33

53 International entrepreneurship builds upon most of the above-mentioned approaches to internationalisation. For example, entrepreneurship is regarded as a unique resource capable of identifying opportunities as they emerge and obtaining needed resources to make use of these opportunities (Ruzzier et al, 2006; Alvarez and Busenitz, 2001;Penrose, 1959). Entrepreneurs have unique relationships, experience and capabilities that are complex and hard to imitate and copy (Alvarez and Busenitz, 2001; Ruzzier et al., 2006). They are agents of change who carry out entrepreneurial and innovative actions and tend to quickly scan the markets for opportunities (exploration) and exploit business opportunities in markets through forming cooperative ventures in foreign markets (Andersson, 2000; Ruzzier et al., 2006). International new ventures: a challenge to the traditional stage and process models of internationalisation Following the pioneering contributions of Oviatt and McDougall (1994; 1995; 1996; 1997), many studies have tried to explain the behaviour of NV using different perspectives and theories such as network theory and organizational learning perspectives. All of these have offered key insights into the phenomenon of NV rapid internationalisation. However, empirical results have shown a mix of some commonalities and some differences. Most of these studies argue that NVs follow a different pattern to internationalisation than traditional large MNCs, where the latter follow a slow and incremental internationalisation pattern and increase their level of internationalisation as more experiential knowledge about the markets is gained and accumulated. Traditional MNCs have mostly started within their domestic home country and then after years of operation expanded to various markets and increased their level of commitment (Aspelund and Moen, 2001; McDougall et al., 2003; Moen, 2002; Servais & Rasmussen, 2000; Rialp et al., 2005), whereas NVs are more capable of internationalising at inception or soon afterwards (Rialp et al., 2005; Knight and Cavusgil, 1996; McDougall et al., 1994; Moen, 2002; Oviatt & McDougall, 1997). 34

54 The most commonly used definition of new ventures is the one offered by Oviatt and McDougall (1994:72): "a business organisation that seeks from inception to create competitive advantage through use of its key resources and sell its products/services in multiple countries ". The age of the firm is a crucial feature of that definition, they argue if the firm internationalises within the first six years of its inception, this means that this took place within the formation stage of NV. There are other factors that contribute to the accelerated internationalisation of NVs, the most critical one being entrepreneurship behaviour. International entrepreneurship involves a "combination of innovative, risk taking behaviour that crosses national borders and is intended to create value for the organisation" (McDougal and Oviatt, 2000:903). Thus NVs tend to have founders with an international orientation which is often translated into entrepreneurial behaviour that crosses national borders. This is achieved through their unique capability to scan the markets and identify opportunities as they emerge. What helps and makes one entrepreneur different than the other is her or his previous experience and a cognitive structure that evolves over time. Their cognitive structural frames of mind as well as prior experience affect the strategic choices that founders make (Arenius, 2002; Shane, 2000). Other factors have been also highlighted such as being embedded in local networks; whether these are industrial zones, scientific networks or networks of small manufacturers (Uzzi, 1997; 1999). 35

55 Different perspectives have been used to understand the behaviour of NVs, such as a knowledge-based view of the firms and network perspectives (Madsen & Servais, 1997; Zahra, Ireland, & Hitt, 2000; Zahra, Matherne & Carleton, 2003; Rialp et al., 2005). Also, the role of the entrepreneurs or managers of NVs in pursuing international strategies (especially if they have had previous international experience, which equips them with the necessary skills and competencies) to enter new markets has been argued to be one of key explanatory factors of rapid internationalisation of NVs (Ellis and Pecotich 2000; Bell, McNaughton, Young & Crick, 2003). Finally the role of serendipity in creating opportunities to enter new markets has also been addressed as a contingency (Crick and Spence, 2005). Other factors contributing to the accelerated internationalisation of those firms have been reported by Preece et al. (1999) and Bloodgood, Sapienza, and Almeida (1996). The former posited that NVs which focus on a narrow market niche and have a short product life cycle will only survive if they seek to internationalise soon after their establishment. The latter proposed that NVs seek to internationalise rapidly if they cannot survive domestic competition and can make more efficient use of their competencies and capabilities in foreign markets. However, other divergent studies have found that born globals are not restricted to those companies coming from high-tech industries and/or new sectors (Bell et al., 2003; Rialp, Rialp and Knight, 2005) but rather they are found in other sectors, such as services (Madsen, Rasmussen & Servais, 2000; McDougall et al., 2003; Moen, 2002; A. Rialp et al., 2005). One plausible reason for these divergent empirical results suggested by Madsen and Servais (1997) is that born globals originating from large countries with large domestic markets (such 36

56 as the USA) operate in high technology intensive industries, while born globals originating from small countries such as Denmark and Norway, operate in other sectors that are not necessarily high-tech. One of the key reasons behind the NVs success and rapid internationalisation, as addressed by a large number of studies, depends to a great extent on the firms' contacts and relationships with specific networks that exist in the new markets. These networks are established and nourished by the firms' founders and key actors' interpersonal relationships with members of other networks. These contacts have been argued to offer access to needed market knowledge and complementary resources that a NV does not have (Harris and Wheeler, 2005; Ojala, 2009). This means that the quality of the relationships and the effort exerted to keep and sustain such relationships play a crucial role in the success or failure of NVs in such markets. These relationships are said to reduce the impact of psychic distance, liability of foreignness and shortage of resources (Harris and Wheeler, 2005; Rugman and Verbeke, 2007; Ojala, 2009). In addition, serendipity plays a role whereby a valuable opportunity arises without planning (for instance, the entrepreneur meets up with one of key distributors in an overseas trade fair and this serendipitous relationship could evolve into a valuable business relationship) which creates chances for SMEs to enter new markets (Crick & Spence, 2005). In an attempt to gather together the various perspectives and factors explaining the behaviour of NVs, Oviatt and McDougall (1994) formulated an integrative model that builds on the key aspects of the classical MNC internationalisation models and applied these to internationalisation behaviour of NVs (Fletcher, 2007). These classical theories were transaction cost analysis, the resource and knowledge-based views of the firm, international 37

57 entrepreneurship, and network theory (Barney, 1991). Their model focused on three main pillars: (1) the characteristics of the NV s founder (2) the NV s founder(s) or management team s early internationalisation motives and (3) the structural mechanisms that firms use to seize international opportunities as they emerge (c.f. Arenius, 2002). Oviatt and McDougall s (1994) model of NV s internationalisation is composed of four main dimensions namely internalization of some transactions, alternative governance structures, foreign location advantage and unique resources. Each of these elements is described in the following section. Oviatt and McDougall s (1994) new venture typology Firstly, firms internalise some of their transactions during their start up stages 5. This is based on transaction cost theory, which argues that firms are formed when market imperfections exist (Williamson, 1985) and that MNCs internationalisation depends on choosing an entry mode that selects the least costly location to operate and involves minimal transaction costs (Buckley and Casson, 1976; Fletcher, 2007). Transaction costs are incurred when the firm devotes a significant amount of time and effort in organizing and controlling transactions across borders. Firms might find it cheaper to internalize these transactions than organizing and carrying out these cross-border transactions at arm-length (Fletcher, 2007). By doing so, the firm accumulates and owns key valuable resources that serve as the basis for future economic transactions. In the case of traditional MNCs these foreign transaction costs are minimized through FDI in the foreign markets. Therefore, Oviatt and McDougall (1994) argue that although "ownership of foreign assets" is not a key distinguishing feature of NVs, both MNCs and NVs need to have ownership of key 5 In their original article Oviatt and McDougall (1994) used the term international new ventures [INVs].In this chapter I will use just new ventures [NVs]. 38

58 valuable assets that serves as a source of economic exchange. Therefore, NVs need in their early start-up stage to have some key assets (accumulated assets) that are internal to the firm and serve as the basis for economic transactions. Secondly, newly established firms suffer from limited resources and power. They tend to "own or internalize" (Oviatt and McDougall, 1994:53) only a limited number of key valuable assets or competencies that are essential for their survival and rely on other means of acquiring those assets that they lack. Therefore, NVs rely on alternative control and governance mechanisms such as network ties, licensing, franchising and alliances with other partners to acquire and have access to those resources needed in order to succeed in new markets. This is argued to be the key distinguishing feature of NVs as opposed to traditional large MNCs. Although, there is always the risk of opportunistic behaviour by one of the partners or an appropriation of NV proprietary knowledge and other competences, NVs still need to rely on these collaborative agreements to compensate for their poverty of resources (Oviatt and McDougall, 1994). Social networks and their inherent informal aspects play a significant role in reducing this opportunistic behaviour (such as trust and cooperation). Thirdly, locating the operations of the firm in a foreign market is what differentiates an international firm from a domestic one. Firms strive to locate transferable resources (such as knowledge and raw materials) in those locations where they can combine non-transferable resources (such as market and location specific resources). This provides MNCs with an advantage over local competitors who lack such transferable resources. However, operating 39

59 in a foreign market creates some challenges for the international firm as opposed to domestic competitors, such as non-tariff trade barriers that are imposed by local governments, language and cultural barriers. Here large MNCs have the advantage of economies of scale that the NV lacks. Accordingly, NVs rely on other resources and competencies to overcome such challenges. One of the key resources is the NV's proprietary knowledge, which is easily transferred, reproduced and combined with other non-transferable resources in many locations. Private knowledge can create either differentiation or cost advantages for the NV in foreign markets over their domestic rivals. This can even occur concurrently in several locations (Oviatt and McDougall, 1994). This provides an explanation of why knowledge intensive industry internationalises rapidly, and more specifically why knowledge intensive NVs internationalise directly or shortly after their inception (Oviatt and McDougall, 1994; Fletcher, 2007). Given the recent increase in the efficiency of international markets, the competition process became more rapid. Thus, for a NV to succeed, it has to internationalise directly after inception. The first three elements namely ownership of foreign assets, ownership or acquiring resources through other means of governance and location decisions set the foundation for a NV to enter and operate in foreign markets. However, for a NV to compete and sustain a competitive advantage it has to have a unique set of resources that are imitable, unique and hard to copy by their rivals (Barney, 1991). Therefore, Oviatt and McDougall (1994) claim that the fourth element, namely unique resources, is vital for the NV s sustainable 40

60 competitive advantage. After a while, the knowledge-base of the firm can be easily copied and disseminated to others. This intimidates the NV s "rent-earning opportunities" as the knowledge is no longer firm-specific and unique to the NV. Hence its ability "to reproduce and transfer this knowledge at a zero marginal cost is a simultaneously beneficial and troublesome property" (Oviatt and McDougall, 1994:56). NVs can prevent the requisition or takeover of their knowledge resources through use of direct proprietary rights (such as patents and copy rights); creation of imperfect limitability of firms resources (where the firm s rivals find it difficult to copy a product or highly embedded firm-specific capability that has an indirect relationship with the firms competitive advantage, such as history of organization and its management style); and forming licensing, network or strategic alliances, with partners who have complementary capabilities in order to tab in the ventures or network knowledge base and unique resources. Moreover, members of the same network share risks and profits (Oviatt and McDougall, 1994; Fletcher, 2007). It is argued that across border networks and alliances should be built on trust and commitment (mutual gain and reciprocity aspects) so that the opportunistic behaviour and risk of expropriation of the unique resources is minimized. These elements are important for the sustainability of NVs but they are manifested in different ways. Oviatt and McDougall (1994) identified four main types of NVs based on two key dimensions, namely; the number of value-chain activities that are coordinated across countries and the number of countries involved as shown in Figure

61 Coordination of value chain activities Few activities coordinated across countries (mainly logistics) Export/import start-up (1) Multinational trader (2) Many activities coordinated across countries Geographically Focused start-up (3) Global start-up (4) Few Many Number of countries involved *Source: adapted from (Oviatt and McDougall, 1994; Porter, 1985) Figure 2.2: Types of new ventures 1. Export/import start-up: These are NVs with few coordinated activities across countries and are involved in few countries. 2. Multinational trader: These NVs are involved in many countries with only few coordinated activities across these countries. 3. Geographically focused start-up: These are NVs that have many coordinated activities across countries but are involved in few countries. 42

62 4. Global start-up: These are NVs which are involved in many countries and have many coordinated activities across these countries. Both export/import start-ups and the multinational traders are called new international market makers, where they tend to focus on coordinating a small number of important value chain activities mainly systems and knowledge found in inbound and outbound logistics, and coordinate the rest of the value chain activities through other governance mechanisms such as networks and alliances. These international market makers keep foreign direct investment [FDI] to a minimum and tend to uncover "imbalance of resources between countries" and create market opportunities that were not there before" (Oviatt and McDougall, 1994:58). Moreover, they tend to achieve a sustainable competitive advantage through their ability to develop and maintain key reliable networks relationships and their exclusive ability to scan various markets and identify those opportunities and have a first mover advantage. The multinational player tends to scan markets, identify opportunities fast and establish needed business networks fast enough before its rivals come into the markets and reap the benefits of the first mover. The geographically concentrated start-ups focus on satisfying the needs of specific regions of the world and coordinate more value-chain activities across countries than the export/import start-up. In addition to inbound and outbound logistics, they coordinate imitable resourcesthat are highly tacit and complex in nature such as human resources and technological development- and might benefit further through forming alliances and networks with other partners (Oviatt and McDougall, 1994). The global start-ups tend to be the most complex type of NVs. They tend to have no geographical constraints and focus on their "proprietary networks" to continuously access and know about the market opportunities as they arise (Oviatt and McDougall, 1994: 60). 43

63 Figure 2.2 presented four "ideal type" categories of NV along two main dimensions. However, Oviatt and McDougall (1994) argue that there are mixed types of NVs between these extremes. Furthermore, this model of NVs is rather static and NVs could change over time through entering additional or fewer foreign countries and coordinating to varying degrees value chain activities across countries. Based on the above discussion, firms need to have ownership of key valuable resources, depend on alliances and networks to secure access and control over other vital resources, such as market knowledge, and also possess unique resources and competencies that are transferable or reproduced in other foreign locations. Accordingly, NVs benefit from first mover advantages and sustain a competitive position. However, Oviatt and McDougall s (1994) model has been criticized because it focuses mainly on knowledge-based NVs and adopts a static rather than a dynamic view. In other words, they focused on the first internationalisation attempt(s). This criticism has led to new ways of classifying NVs (Bell and McNaughton, 2000; Oviatt and McDougall, 2005b) Other SME typologies Bell and McNaughton (2000) classified NVs into ''knowledge intensive and knowledge-based firms" (Rialp et al., 2005: 159). The former refers to those firms who rely extensively on acquired knowledge in developing their products, services and improving their productivity and innovativeness but do not rely on creating new knowledge as their basis for operations 44

64 and processes (Rialp et al., 2005). Software-writing firms tend to fall within this category. In other words, they have an exploitation-based competitive advantage (March, 1991). By contrast, "knowledge-based" refers to those firms that emerge as a result of new technological developments and new sectors in the economy (such as biotechnology) and the complex knowledge inherent in the company processes. This is the basis of their core competencies and they have a so-called exploration-based competitive advantage (Rialp et al., 2005). Bell, McNaughton, Young and Crick (2003) suggested a similar typology but added another type of firms which they referred to as "traditional firms". Clothing firms offer an example of the traditional type. They classified SMEs based on variations in novelty, complexity and sophistication of knowledge which are the key drivers behind the speed of an SME s internationalisation. Traditional firms are firms that tend to bring a well-established technology into a new context or market (Oviatt and McDougall, 2005b:543). Accordingly, countries with diverse technological sectors would experience the emergence of NVs that vary in the knowledge and service intensity of their operations. In other words, NVs are not restricted to those companies operating in high-tech knowledge-based sectors but rather they may prevail as well in other sectors. Child and Hsieh (2014) suggest that these three types of SME traditional, knowledge-intensive and knowledge-based will tend to vary in their decision making on internationalization as well as in their use of information and network attachments. 45

65 The internationalisation process (or speed of internationalisation) of these three types is likely to be different. Traditional firms tend to follow the traditional Uppsala model whereby the firm internationalises gradually and slowly. On the other hand, the other two types tend to have a faster and accelerated rate of internationalisation, whereby the knowledge-intensive firms tend to internationalise faster as their key competency can be transferred and used in several countries. Knowledge-based NVs tend to internationalise very fast as their whole existence depends on "new complex knowledge" which can be deployed and sustained in multiple locations (Oviatt and McDougall, 2005:543). The characteristics of knowledgebased NVs are best explained by Oviatt and McDougall s (1994) integrative model. To sum up, most studies have ignored the mode of entry that is employed by different types of SMEs and how it changes or evolves as the SME becomes more mature and well established in foreign markets. Thus, knowledge-based firms might start off with licensing or forming alliances with local partners in foreign markets then later increase their market commitment as they develop a larger customer base and need to have most of their value chain activities coordinated and carried out in foreign markets. This could be done through having a green-field subsidiary to be near their customers for example, to offer after-sales services (Hashai and Almor, 2004). Also, a knowledge-based SME might start off by exporting some of its key products or services or becoming a supplier of a key component to large multinationals in various markets. The internationalisation process of SMEs deserves further empirical investigation aimed at uncovering some of its complexities (Fletcher, 2007; Oviatt and McDougall, 2005a; Autio, Sapienza and Almeida, 2000). SMEs ''have different routes and follow different patterns of 46

66 internationalisation depending on the level of knowledge acquired prior to the birth of a NV and the existing and new networks'' that offer access and complementary resources" to the firm (Christensen and Jacobsen, 1996:7; Madsen and Servais, 1997). The SME's network ties and relationships in foreign markets need to be taken into account as do other potential influences including national context, sector and previous experience. Summary of internationalisation models and perspectives Ahokangas (1998) posited that SMEs internationalisation models can be viewed from three different angles: the market, firm or entrepreneurship perspectives. The market perspective has focused mainly on MNCs and their diversification strategies that are deeply rooted in economics (Dunning, 1988). The firm perspective has been the focus of stage models of internationalisation and the entrepreneurship perspective has only recently emerged. Both innovation models and Uppsala model have been applied to small and large firms, arguing that firms follow an incremental process of internationalisation and they increase their level of commitment and involvement as their level of experiential learning increases (Ruzzier et al., 2006). The network perspective and resources based view are closely connected. Both perspectives focus on the internal and external resources (obtained through the firm s network) that the firm needs in order to successfully internationalise. A firm can cooperate vertically with other firms at different stages of production. For instance, it could cooperate with a supplier and with a distributor. It could also collaborate with its competitors through forming horizontal network relationships (Ahokangas, 1998; Ruzzier et al., 2006). There is also a 47

67 close link between entrepreneurship and network perspectives. An important component of entrepreneurship is the creation of social capital through forging network links which assist internationalisation. Based on these considerations, networks and the social capital they create are important for the survival and growth of SMEs, not least in terms of their internationalisation. The following section provides detailed background of networks and social capital theories Theories of social capital The concept of social capital is not a new one. It has been used by academics since the early 1900s. It is deeply rooted in the work of sociologists such as Bourdieu (1986) and Coleman (1988). The former was the first to offer a detailed definition of social capital: "it is the aggregate of the actual or potential resources which are linked to possession of durable networks of more or less institutionalized relationships of mutual acquaintance or recognition" (Bourdieu 1986: 248; Portes, 1998). Coleman (1988; 1990) on the other hand offers another comprehensive definition of social capital that has been widely used by scholars in the field. He argues that social capital is defined by its function. It is not a single entity, but a variety of different entities having two characteristics in common. They all consist of some aspects of social structure, and they facilitate certain actions of individuals who are within the structure. Like other forms of capital, "social capital is productive, making possible the achievement of certain ends that would not be attainable in its absence" (Coleman, 1990: 302). Another key sociologist who took a different view of social capital is Burt (1992). He introduced the concept of structural holes, whereby an actor is said to be occupying a structural hole if he or she is connected to a group of unconnected actors. The actor then 48

68 occupies a brokerage position that gives him control over the flow of information and resources and the option of building bridges between groups of disconnected actors whenever needed (Burt, 1997). In later publications, Burt (2008) argued that being connected to a broker provides access to structural holes which provides the actor with new sources of information and resources that he does not have access to through his/her own network. Burt s (1992) work is built upon the concept of weak ties (Granovetter, 1973; 1974) and the power that sole partners possess (Freeman, 1970; 1979; Burt, 1997). In this case, the network is sparse - consisting of a few redundant ties - resulting in new sources of information and other resources, creating social capital for the actors who have access to structural holes. This view is contrary to the view adopted by Coleman (1988), who argues that closed networks (which are dominated by strong ties) provides access to crucial sources of information. Both these authors, however, have highlighted the fact that actors (whether individuals, groups or organizations) obtain and have access to resources through their external relationships (Burt, 1992; Coleman, 1988). However, the application of the social capital concept has extended beyond the field of sociology. There have been studies of its impact on career success (Burt, 1992; Burt, 1992; Gabbay & Zuckerman, 1998; Adler and Kwon, 2002), how it creates a pool of candidates for recruitment purposes (Fernandez, Castilla, & Moore, 2000), and how it facilitates resource and knowledge exchange between different business units (Gabbay& Zuckerman, 1998; Hansen, 1999; Tsai & Ghoshal, 1998). Other applications concern how entrepreneurs use their own networks or social capital to seize opportunities (Ellis, 2010; Chong & Gibbons, 1997), how social capital facilitates a new venture s establishment (Coviello, 2006; Coviello 49

69 and Munro, 1995; Walker, Kogut, & Shan, 1997) and how it assists inter-firm learning and reduces the likelihood of a firm s disbanding and termination (Pennings, Lee and Witteloostuijn, 1998; Adler and Kwon, 2002). The complexity of social relationships and their different types has been highlighted by Adler and Kwon (2002). They argue that one type of relationship (for example, a former colleague and friend) could serve other purposes (such as offering emotional support and nonprofessional advice). This has been referred to by Coleman as the "appropriability" of social structure (1988: 108). In other words, gathering under one concept are several related theories that have been addressed by scholars over the years such as social exchange theory, trust, role of informal social structure, structural and relational embeddedness and social networks (Alder and Kwon, 2002). Hirsch & Levin (1999) have referred to the use of social capital to cover several theories and perspectives as an "umbrella concept". This has led to several criticisms of the use of social capital as an all-encompassing concept (Lappe & Du Bois, 1997; Narayan & Pritchett, 1997; Adler and Kwon, 2002). Accordingly, Adler and Kwon (2002) argue that there is a need to understand the sources of social capital, the benefits accrued from it, and its risks and contingencies. With this in mind, the following section discusses the different definitions of the concept, its sources and dimensions, different theoretical perspectives underpinning the concept, and the benefits and costs of social capital. Social capital definitions and perspectives There is much debate among scholars on the way they define social capital. Some definitions have emphasized certain aspects of social capital, while other scholars have attempted to find 50

70 a general and holistic definition of social capital (Nahapiet and Ghoshal, 1998; Adler and Kwon, 2002). Several definitions have been used by different researchers. For example, Bourdieu & Wacquant (1992) defined social capital as "the sum of resources that an individual or a group will accrue as a result of being in a network of relationships" (p.119). Hitt and Ireland (2002) emphasized that an organization s and an individual s "set of relationships creates value and facilitates action"(p.5). However, these authors have focused on the individual's perspective and regarded social capital as a private good that benefits only the person who possesses it. On the other hand, social capital can benefit groups of people and even society as a whole (Putman, 1993). This has been reflected in recent studies that have viewed social capital from an organizational perspective. Social capital plays the role of a public good, where it is an organizational resource that benefits all the members of an organization and consists of a set of social relationships within an organization (Ratten and Suseno, 2006). It facilitates the spread of information and at the societal level it is a fruitful division of labour. Social capital has been further viewed from two different perspectives; egocentric /bridging; and socio-centric/bonding. The former emphasizes that individual actor s social capital is characterized by her/his direct and indirect relationships with others in a social network (Ratten & Suseno, 2006). The distinction between bonding and bridging social capital is unclear and sometimes confusing (Prashantham, 2005; Arenius, 2002). Bonding social capital is concerned with bringing together people who are similar to one another in one or more aspects (such as ethnicity, age, gender and social class). By contrast, bridging social capital is concerned with bringing together people who are dissimilar (Putnam and Goss, 2002). In 51

71 other words, bonding social capital is represented in homogenous groups (sharing the same demographics or some common characteristics, for instance relationship with parents or kinship) and bridging social capital is represented in heterogeneous groups (relationship with social milieu) (Putman, 2000). One distinction between bonding and bridging social capital that has been proposed by Adler and Kwon (2002) is that bonding social capital refers to intra-firm relationships whereas bridging social capital resides outside the organization (inter-organizational linkages or external relationships). However, caution must be applied when using the latter distinction as both external and internal relationships tend to overlap. 6 Bridging social capital helps in understanding why some individuals are more successful than others and why some firms are more competitive than others. As noted earlier, in Burt's (1992) view an individual can have access to valuable information due to brokering opportunities inherent in his/her location in a network of structural holes (where s/he is a focal actor who is associated with a mediator between two unconnected groups). This means that individual social capital plays the role of a private good or asset that benefits the actor or individual or organization who possesses it and is often reflected in his/her personal success (Ratten & Suseno, 2006). Bonding social capital focuses on "the social structure of the collective actors, whether these actors are groups, organizations or communities rather than individual actors per se" (Ratten & Suseno, 2006:62). This is evident in Colman s (1988) view that closed networks (usually characterized by strong ties) are a vital source of 6 The distinction between bonding and bridging social capital in this research will follow that of Putman and Goss (2002). 52

72 information and resources. In the context of an SME s social capital can often take the form of close personal bonds. Despite the differences between Burt (1992) and Colman (1988), it is clear that resources could be obtained and accessed via the network relationships of an actor or an organization (Prashantham, 2005). Therefore, the presence of an actor s set of relationships creates an opportunity for this actor as opposed to its absence (Arenius, 2002; Burt, 1997). 7 Given the different conceptualizations of definitions of social capital, three important considerations need to be addressed. Firstly, social capital consists of two aspects which are the networks of relationships and the resources that are accessed through the presence of these relationships (Arenius, 2002). Secondly, the divergent views and classification of social capital as a private versus public good and the emphasis on internal social capital (bonding social capital) versus external social capital (bridging social capital) needs to be incorporated in an integrative definition of social capital (Adler and Kwon, 2002). Thirdly, the distinction between the private versus public good aspects of social capital needs to be incorporated in the definition 8. Accordingly, the working definition of social capital used in the present study is adapted from Nahapiet and Ghoshal (1998): "social capital is the sum of actual and potential resources that are embedded within and made available or derived from a network of relationships possessed by an actor or a social unit" (1998:243). Thus it is the individual actor 7 Social capital is different from intellectual capital which focuses on the individuals (or organization s) abilities and skills and is different from human capital which refers to stock of employees that an organization have and others do not have (Adler and Kwon, 2002). 8 The distinction between public and private good of social capital is not used in this study as the focus is primarily on the private good view. In other words, it is concerned with the social capital of the SME decision maker that makes his more (un)successful compared to other SME decision makers from the same context. 53

73 (entrepreneur) or social unit s (SME) network ties that provides "access to valuable resources" and information (Nahapiet & Ghoshal, 1998:252). This could involve relationships with family members, friends, government agencies, clients, competitors, business partners, suppliers and distributors that provide the entrepreneurial firm with access to the resources needed over time and helps the firm to survive and grow. It involves both bonding (for example family and friends) and bridging social capital. Nahapiet and Ghoshal (1998) have identified three main dimensions of social capital namely, structural, relational and cognitive dimensions. The structural dimension refers to actual configurations of the network, number of ties and other structural aspects. The relational dimension refers to the nature of relationships and the resources that could be obtained from these relationships such as trust and cohesion, identification and respect (Tsai and Ghoshal, 1998). The cognitive dimension refers to the set of resources that provides common "meaning and understanding to network members "such as shared goals, norms, common understanding, shared language" (Inkpen & Tsang, 2005:158; Nahapiet and Ghoshal (1998:251). Several authors have employed different aspects and facets of social capital in their analyses. The most important distinction highlighted by Granovetter (1973) is between the structure and the quality of ties (Arenius, 2002). The quality of ties could be classified into weak and strong ties. The former refers to relationships that are characterized by a limited frequency of interaction, while the latter refers to the relationships that are characterized by high frequency of social interaction (Woolcock, 1998; Granovetter, 1973). However, certain ties could be 54

74 characterised as a weak tie even with high frequency of interaction if it is merely contractually based and both parties meet only to discuss business related matters. The quality of relationships incorporates both the relational and cognitive dimensions of social capital. The present study focuses on both the structural (network) aspect of SC and the quality of relationships which includes the relational aspect only. The cognitive dimension is excluded because this study focuses on firm-level social capital (using an egonet perspective) rather than a dyadic view (which would be necessary to assess if both parties have the same shared understanding and norms). Benefits and costs of social capital Given the above conceptualizations of social capital, several benefits can be obtained from individual and organizational social capital (in most cases they tend to overlap). Organizational member relationships can help the company to seize opportunities in the market without participating in obtaining or creating them. Social capital provides informational, solidarity and influence benefits (Adler and Kwon, 2002). Market knowledge is vital to conduct business in foreign markets and to access new and diverse information and referrals (Gulati, 1999; Burt, 1992). Social capital also enhances the legitimacy and goodwill of actors (Burt, 1992; Arenius, 2002). However, social capital is not without costs and negative aspects. There is the cost and time involved in maintaining relationships (Prasthantham and Dhanaraj. 2010). There is also the risk of closure, which means that the focal actor becomes blinkered and enclosed within his/her closely knit network and becomes unable to recognize or access new sources of information (Coleman, 1988; Adler and Kwon, 2002). For example, members of the same 55

75 rotary club share the same jokes and information day after day and they sometimes become so focused on themselves that they do not mix with other members outside their social group and do not hear about new jokes and information. These closely knit networks serve as a source of redundant ties (Burt, 1992). In addition, there are risks associated with uncertainty regarding how business relationships will evolve, especially if the focal firm has limited market knowledge and the local partner is "powerful enough to exercise leverage within that location" (Child and Rodrigues, 2004:96). Another set of costs would involve the leakage or learning race referred to by Hamel (1991), whereby a focal firm s key knowledge assets (Boisot, 1998) and resources are acquired by the local partner and this partner develops a superior learning ability than the other firm and may even find it unattractive to maintain the relationship (Child and Rodrigues, 2004). Some studies have used the terms social networks and sometimes social capital interchangeably. In her study Coviello (2006) did not make a clear distinction between networks and social capital. It can be argued that a network by itself does not necessarily convey social capital. Social capital is only created when benefits or value can be extracted from a network of relationships. The following section discusses networks as a key component of social capital, followed by how both the structural and relational dimensions of social capital evolve over time. 56

76 Networks and social capital Networking means that an actor connects with one or more actors within a given collectivity in order to gain access to needed resources. The connection between actors can be viewed from various levels of analysis. The micro perspective focuses on the dyadic level which is the connection between two actors (Hanneman and Riddle, 1998), whereas the macro perspective focuses on the overall network. The former is the smallest social unit and is characterised by the presence or absence of ties (between a pair of actors). The latter the whole network is composed of a set of dyadic structures between actors (Hanneman and Riddle, 1998). There are three possibilities regarding the tie formation: no tie actually exists between actors; there is a uni-directional tie; or there is a reciprocated tie. These possibilities indicate the degree of cohesion and social capital that may be present (Hanneman and Riddle, 1998). In addition, dyads are said to be embedded in local social structures or sub-structures with which they share a common attribute or membership. This is called embeddeness in a local social structure (Hanneman and Riddle, 1998). The purpose of my study is to understand the variation of the behaviour of the entrepreneurial firms as the focal point of analysis. The main premise of ego-networks is to understand how individual actors are embedded in their local structures and how this is reflected in variation of their behaviour (Hanneman and Riddle, 1998). An ego network perspective will be used in this study in order to understand and note the variation across focal firms in terms of how they are embedded in their local context and how this is reflected in their differential internationalisation performance. 57

77 An ego-network revolves around a particular person or organization (Arenius, 2002). This individual or firm is referred to as the "ego" and the persons or firms to which he/she is connected are often referred to as "alters" (Hanneman and Riddle, 1998). For the sake of simplicity, the entrepreneur s set of relationships will be referred to as entrepreneurial networks and the entrepreneurial firm's set of relationships will be referred to as organizational networks. Uzzi (1999) differentiated between embedded and instrumental ties. The former refer to" the degree to which commercial transactions take place through social relations and networks of relations that use exchange protocols associated with social, non-commercial attachments to govern business dealings" (Uzzi, 1999:482). These ties are private and personal in nature as opposed to the instrumental or professional ties which are found in the task or business environment in which the firm operates (Arenius, 2002). However, instrumental ties are arm's length ties similar to market/business ties. This distinction is closely linked to the strength of the tie proposed by Granovetter (1973). The tie is weak if the actors interact infrequently, and is strong when actors interact frequently. Strong ties could exist between friends and family members (Granovetter, 1973). However, as mentioned earlier, frequency of interaction is not necessarily the only indicator of a weak tie. Partners could have weak ties that involves frequent interactions primarily focused on business related aspects rather than a personal or social reason. Ego networks The egonet consists of a set of dyadic relationships between actors and different alters (Arenius, 2002). For instance, assume that company X is the focal actor (ego) in this case and 58

78 has a relationship with actors B and C respectively, the ego s network could be visualized as in Figure 2.3. These ties could be embedded ones (strong personal tie) or weak impersonal/arm's length ties. In addition, the ties could serve more than one purpose and ego s net could be composed of multiplex set of ties. This is depicted by the density of ties (lines) in Figure 2.3. A B C Figure 2.3: An example of ego net (adapted from Arenius, 2002) The blue circle refers to ego (focal firm) and the red circle refers to the alters (A, B and C) with whom the focal firm has a direct contact. Each of these alters are connected to other networks such that the focal firm is only linked to through the main alter. In addition, the red and blue circles refer to the first order network which is composed of ego s alters. Alters connections (boxes) refer to second order network (Arenius 2002; Hanneman and Riddle, 1998, Burt, 2010). In addition each one of those alters alters could be connected to one another. However, this last possibility is beyond the scope of the present study. In order to understand how the network evolves and the processes underlying its evolution, it is important to understand what happens at the individual dyadic level (if possible) and then with the accumulation of dyads into a broader network. Thus the evolution of a network 59

79 depends on the evolution of the individual dyads that constitutes the overall network (Hanneman and Riddle, 1998; Arenius, 2002, Slotte-Kock and Coviello, 2010). Accordingly, it is important to focus on: (1) General network measures of the overall ego network such as network density and size (Coviello, 2006; Arenius, 2002). (2) Structural hole measures such as network constraint (Burt, 1992; Coviello,2006) The first measure is network size which refers to the number of contacts present in ego s network; this reflects the degree of involvement of the actor in relationships (Arenius, 2002; Coviello, 2006). In other words, the higher the ego s network size the more likely that the ego will have more opportunities and options, and hence greater social capital. This creates a degree of self-sufficiency whereby the actor becomes less dependent on a restricted number of actors. It also creates more power to the ego (Arenius, 2002). The second general measure is network density, which refers to the percentage of total ties that are connected to one another taking into consideration the number of potential pairs of ties (Borgatti, Jones and Everett, 1998; Coviello, 2006). The impact of high density networks is debated. One line of thought is associated with Coleman (1988) who argues that network closure (high density) is characterized by a higher degree of trust and reciprocal interests providing better sources of information. A second line of thought is associated with Burt s (1992) notion of structural hole, and argues that the higher the density of ego s network the more likely there will be redundant ties and less social capital. In other words, these redundant ties only offer redundant sources of 60

80 information and resources. If the actor (ego) has access to structural holes or connects between unconnected actors he is in a better position and thus has higher level of social capital (Burt, 1992; 2005). These views appear to be contradictory. One favours closure as opposed to a sparser network. I shall assume in this thesis is that they are complementary, and that it is functional to have closely connected ties characterized by high degree of trust as well as some weak (nonredundant ties) to serve as new sources of information (Burt, 2000; Coviello, 2006; Han, 2006) This leads to the second set of measures which are the structural holes measures. The key requirement here is to minimize the number of redundant ties, which will create more entrepreneurial opportunities to the SME (Greve, 1995). Following Coviello (2006) and Borgatti et al. (1998), network constraint will be used as a measure of structural holes. It measures the extent to which ego (the SME) has invested (is connected to) in a single group of actors (Burt, 2009). In other words, it refers to the extent to which all ego s (the SME's) ties directly or indirectly involve only a single actor (Borgatti, Everett and Freeman, 2002; Coviello, 2006). This means that there are more opportunities to act for the ego, if there is a low degree of constraint (Coviello, 2006). The role of social capital in facilitating SME internationalisation and leading to improved performance is amply recognized in the literature. However, the ways through which social capital develops over time are not clearly understood (Prashantham and Dhanaraj, 2010). 61

81 Specifically, the processes that enable social capital to develop and those that constrain its development and how this balance is reflected in the firm's performance are unclear (Mauer and Ebers, 2006). Social capital could escalate or depreciate like most forms of assets, making it important to understand its development and evolution over time (Prashantham and Dhanaraj, 2010:967). It has been shown that social capital is important for the firm s international expansion. However, little is known whether over time social capital will be sufficient for successful growth and survival of an entrepreneurial firm. Relationships do not remain in their current form, but rather they tend to appreciate and depreciate over time. Some relationships (such as business relationships) are restricted to a given economic purpose, and once their utility expires, the tie decays (Arino, Ragozzino and Reuer, 2008; Doz, 1996; Ring and Van de Ven, 1994; Prashantham and Dhanaraj, 2010). Thus as the entrepreneurial firm evolves and develops, the significance of different types of relationships will differ as it moves from one developmental stage to another (Coviello, 2006). This implies that the firm s social capital evolves as it moves from one stage of internationalisation to another. The following section discusses the evolution of social capital. Evolution of social capital The evolution of social capital involves the evolution of the overall network and the individual dyads that make up the firm's ego net. Building on the international entrepreneurship literature, some useful insights can be drawn for conceptualizing the dynamics of network evolution. Larson and Starr (1993) and Hite and Hesterly (2001) have advanced two differing views on how the networks of small entrepreneurial firms evolve over 62

82 time. The former argue that the firm starts with an uncomplicated, one-dimensional set of dyadic relationships that are dense and close, moving subsequently into a set of multiplex and multifaceted organizational relationships. They argue that usually the firm starts off with a set of largely but not exclusively social relationships (friends/family or previous business connections), then in later stages they are converted into combination of both business and social relationships. This adds a compound effect on the firm resulting in a more solid integration of dyadic relationships and an increased number of business ties (Coviello, 2006). On the other hand, Hite and Hesterly (2001) claim that firms start off with a set of interconnected relationships (consisting only of social relationships) on which the firm relies heavily for its growth; and as the firm moves into its growth and stability stages its network comes to comprise both "embedded and arm s length" business ties. In other words, the firm's network moves from a self-based, path dependent one to a consciously evaluated and intentionally managed one (Coviello, 2005:715). These two studies have a common base. Both maintain that firms will start off with strongly connected social relationships and will develop in a linear way. However, they differ in that Larson and Star (1993) argue the network will evolve into a denser and highly integrated structure composed of both social and business relationships, which overlap as the SME moves between the early stages of the firm's lifecycle and later stages. Whereas Hite and Hesterly (2001) suggest that firms will move into a sparser, open network, which will be composed of more business ties in the later stages of development. A final point of departure between the two studies is that Larson and Star (1993) implicitly suggest that networks are intentionally managed, whereas Hite and Hesterly (2001) argue that they are not necessarily intentionally managed from conception. 63

83 Although these two studies focus primarily on early-stage entrepreneurial firms, they offer important theoretical conceptualization of how networks evolve over time. Coviello (2006) has examined the evolution of the overall network in terms of its density, size and centrality over time, concluding that the early-stage entrepreneurial firm s network size will increase but their network density will decrease over time and the firm s network will be dominated with business ties during the initial and early growth stages of firm s lifecycle. Prashantham and Dhanaraj (2010) focused on individual ties and how they evolved in terms of expansion versus decay. They argued that the firm starts off with an initial stock of social capital (through a set of ties) and that over time it evolves and creates dynamic social capital. They also explained how this creates network learning crucial for the firm s growth. Building on the above arguments, the present study aims to examine the evolution of an SME's network as a whole and the individual dyads that make up that whole network. This will represent the structural dimension of social capital. It involves focusing on key structural measures (Borgatti et al., 1998; Coviello, 2006) and the patterns of development of individual ties (expansion, contraction), and how and why such patterns occur. Research on networks and the internationalisation of SMEs To date, numerous studies have emphasized that networks have a positive impact on the internationalisation of SMEs and on the speed of that internationalisation. Two major reviews by Hoang and Antoncic (2003) and Slotte-Kock and Coviello (2010) identified three main categories of research on the use of networks in the entrepreneurship literature. These three categories were (1) studies that focused on how networks impact entrepreneurship processes and outcomes (treating networks as an independent variable); (2) how the entrepreneurship process affects network development (treating networks as a dependent 64

84 variable); and (3) studies that treated networks and internationalisation as inter-dependent. The literature addressing networks and SME internationalisation can be grouped into three main streams as shown in Table 2.4. These reviews suggested that future research taking networks to be an independent variable, should use longitudinal studies to understand how networks shape opportunities being pursued. Whereas researchers who view networks as a dependent variable should investigate how inter-organizational relationships are developed at the dyadic level. Both recommendations are embedded in the notion of process within the context of change (Slotte- Kock and Coviello, 2010). Hoang and Antoncic (2003) suggested that researchers should use an integrative approach instead of positioning network as a dependent or independent variable. This is has been evident in studies that have focused on social capital and immigrant entrepreneurs. It represents the third category, treating networks and internationalisation as inter-facing and inter-dependent. 65

85 Table 2.4. Key themes examining the importance of networks in SME internationalisation Main themes/streams of literature Networks as an independent variable Dynamics of networking process through positioning networks as an independent variable Key aspects/ theoretical lens These studies focused on the SME's or entrepreneur s perception of network; whereby they examined how the networks impact the SMEs internationalisation processes and outcomes. For example, some scholars focused on the usefulness and quality of ties, the functions or benefits of relationships (c.f. Harris and Wheeler, 2005; Uzzi, 1997; 1999; Loane and Bell, 2006). Generally, key benefits and functions offered by network relationships to SMEs are not precisely specified nor thoroughly researched. Few studies focused on how networks facilitate internationalisation process of SME over time (Hoang and Antoncic, 2003; Coviello, 2006; Slotte- Kock and Coviello, 2009; Jack, 2005; Coviello, 2006). These studies have addressed this aspect drawing on structural dimensions of networks such as network cohesiveness and reliance on bridging relationships (Coleman, 1988; Burt, 1992), strength of relationships and embeddness of the SME in its social context (Granovetter, 1973; Uzzi, 1997;1999). Or they compared the structural evolution of SME networks over the early stages of the SME life cycle (for example, Coviello, 2006). The structural configuration of networks serves as a proxy for the SME s access to resources and information which are made readily available through its membership of a particular network. Other studies focused on individual network tie characteristics through addressing the question of what is the best combination of ties (weak or strong, business or social) that could create benefits for the company (Bulter and Hansen, 1991; Hite and Hesterly, 2001; Steier and Greenwood, 2000). These studies focused on how SMEs network develop over time through addressing the evolution of the ties from social to business ties Comments Depended mainly on cross-sectional studies to try to understand how networks facilitate the internationalisation of SMEs and/or few case studies (between 3-9 case studies). These studies were either conceptual in nature (e.g., Hite and Hesterly, 2001; Larson and Star, 1993; Han, 2006) and/or they focused on a narrow perspective of network evolution over time through using simple grouping of relationships to social and business relationships; or broad types of networks such as support networks, professional networks; and/or weak and strong ties. 66

86 Networks as a dependent variable and/or multiplex ties and how the strength of ties evolve over time. Finally, others define network change in terms of how, for example, exchange relationships between an SME and its partners ties become dormant or cease to exist (Prashantham and Dhanaraj, 2010; Jack, 2005). A few scholars began to investigate the key benefits offered to SMEs and how these impact on the entrepreneur s ability to spot and seize opportunities in overseas markets (Ellis and Pecotich, 2000; Blomstermo, Eriksson, Lindstrand and Sharma, 2004). There are fewer studies in this category than in the first one. It is concerned with how an SME's network is affected by entrepreneurial action (Hoang and Antoncic, 2003; Slotte-Kock and Coviello, 2010). In other words, how the strategies of SMEs shape and change networks over time. This steam of literature draws heavily upon the industrial networks literature introduced by Johanson and Mattsson (1994) that focused on how companies manage their interorganisational networks or inter-firm networks Hence the focus is on the overall network of relationships. These studies draw attention to the importance of managing important relationships in a way that maximizes benefits for the firm and that reflects a rational economic logic on the part of managers or entrepreneurs (Slotte-Kock and Coviello, 2010). The focus is on overall networks of inter-firm relationships. Moreover, the key assumption is that the ties they contain mainly reflect a rational choice on part of the exchange partners Haong and Antoncic (2003). 67

87 Dynamics of networking process through positioning networks as a dependent variable These studies focused on assessing how exchange partners work together to make best use of resources through collaboration and cooptation. Also, how different partners try to improve their position through gaining preferential access to strategic partners and how partners try to maximize trust as a key means to reduce competitive behaviour and achieve a win-win situation for all actors involved (Hakansson and Snehota, 1995; Slotte-Kock and Coviello, 2010). These studies focused on interaction between exchange partners and how, at the level of dyadic relationships, the whole network could change as a result of adaptation to external influences (Johanson and Mattsson, 1994; Slotte-Kock and Coviello, 2010). Studies that position networks and internationalisation on the same interface and treat them as interdependent The final stream has adapted a network approach or view on internationalisation in a unique way through arguing that SMEs entry to foreign markets is heavily influenced by the contacts and relationships they have in their portfolio of relationships. The internationalisation process should be understood through the process by which SMEs enter into overseas networks. For example, Jansson and Sandberg (2008) introduced the entry node concept whereby the SMEs entry to overseas markets is contingent upon ties in their networks and these ties are the key way through which companies manage their overseas operations and enter new markets. Hence they are equivalent to the entry mode strategies adapted by the company. Others like Hilmersson and Jansson (2012) have examined the process through which SMEs establish an insidership position in overseas networks. These studies focused on either identifying key nodes or viewed internationalisation as process of entering new markets through development of particular types of network. 68

88 The present study will primarily focus on the interface between both internationalisation and network development over time. It will also examine the benefits accruing from the social capital provided by those networks. The SMEs to be studied are located in contrasting institutional environments. The following section discusses the background to institutional theory, including its cultural dimension, and how it applies to my research. 2.3 Historical evolution of institutional theory Institutions have been a major interest of social scientists for over a century. Table 2.5 provides a summary of influential contributions to the development of institutional theory. 69

89 Table 2.5: The development of institutional theory Time period Authors Description Early 1900s (early intuitionalists) Max Weber (1906/1946), Merton (1936). The roots of institutional theory can be traced back to the concept of bureaucracy introduced by Max Weber's (1906/1946) essays in sociology. His work on bureaucracy and behaviour in bureaucracies was further developed by Merton (1936). Merton argued that formal and rigid organisational rules and formal procedures could hinder the achievement of an organization's goals giving rise to bureaucratic dysfunctions (Merton, Ailsa and Hockey, 1952; Scott, 2001) Selznick, 1948; Parsons,1956; Simon, 1945/ 1997) Selznick (1948), building on Merton s work, treated institutions as a variable phenomenon and a process. He argued that organisations are similar to living organisms, they have to adapt to changes in the environment and it is shaped by the choices and actions of its members. He argued that over time an organisation creates a distinctive set of standards and ideals that result in a separate structure or identity. It then could be viewed as a standalone entity and treated as an exogenous variable rather than a process, where institutions change over time (Selznick, 1948; 1957; Scott, 2001). Parsons (1956) added a new dimension to the early definition of institutions, namely the cultural-cognitive dimension. He argued that shared norms and values govern the individual and organisations behaviours and defines the appropriate course of action. In other words, he argued that individual and organisations actions are regulated by a set of normative and cultural values and that if individuals and organisations measure up to these values they would gain legitimacy and have a higher share of the societal resources (Parsons, 1953; [1956]1960; [1934] 1990; Scott, 1995; 2001). This is closely linked to the concept of institutional isomorphism, which is the process by which firms follow common standards and practices that become legitimised as a result of confirming to norms and values set by their institutional context (Kiss and Danis, 2008; Meyer and Rowan, 1977; DiMaggio and Powell, 1983). Herbert Simon [1945] 1997] added the concept of bounded rationality where he argues that individuals' decision making is contingent upon some value premises that guide the decision making process. within the context of organisation organisational routines, targets and rules equip and sets the framework upon which employees make rational decision regarding how to organisational goals. It is important to note that Simon [1945] 1997] challenges mainstream economic thinking that individual are full rational decision makers but argues that individuals have cognitive boundaries that are contingent upon the information available, values and routines present in a particular context. For example, an employee s decision on how to repair a particular machine is governed by the training and procedures set by his supervisor. 70

90 Neoinstitutionalism North, 1989; 1990; March and Olsen, 1984; Moe, 1990; DiMaggio and Powel, 1991; Nelson and Winter, 1982; Thelen and Steinmo, 1992; Moe, 1990) Neo-institutionalism is rooted in the work of political economists and sociologists (such as North, 1989; 1990; March and Olsen, 1984; Moe, 1990; DiMaggio and Powel, 1991). The best-known contributor is North (1990), who applies transaction cost theory and argues that the economic exchange between economic actors is dependent on the institutional context or framework they are embedded. He argues that wider institutional context such as that of the country sets the "rules of the game" (North, 1990:3; Scott, 2001). This has a bearing on the exchange that takes place between people and organisations. These constraints consist of actions that individuals are forbidden to do and the prerequisites that allow individuals and organizations to take on certain activities. This creates the "institutional framework within which any kind of human interaction takes place" (North, 1990: 4). Evolutionary economists- adhering to neo-institutionalism- argue that only those organisations belonging to a particular population will survive if they have the best fitting characterises to the external environment. These characteristics become internalised into the organisation and in the form of routines and procedures that affects future adaption behaviour of the organisation. These routines and procedures are one form of institutionalisation (Nelson and Winter, 1982; Scott, 2001; Child et al., 2013). By contrast, political sociologists have focused mainly on what Scott (2001) referred to as historical institutionalisation and rational choice theorists. The former group argues that regulatory institutions are not separate from the informal institutions that are constructed by the social actors but rather they are interrelated. They argue that institutions are composed of formal and informal systems that govern individual and organisational behaviours (Thelen and Steinmo, 1992; Scott, 1995:2001). They emphasise that the institutional framework could constrain or facilitate the individual actors behaviour. The latter emphasises the importance of the regulatory aspect of institutions and argues that social actors are rational human beings and that they build political and regulatory institutions that help in governing and protecting their interests (Moe, 1990; 1984; Scott, 1995, 2001). These institutions take the form of rules and regulations that governs sanctions and empowers individuals within a particular context. Some organisational sociologists have added a further dimension to the definition of institutions. This is the culturalcognitive dimension whereby symbols and values are seen to form a unique social institution that shapes individual and organisational behaviours. These cultural values are institutionalised through the process of socialisation (social interaction between actors). Over time they become internalised to societies, forming part of the cultural heritage of countries and communities (Scott, 2001; Berger and Luckmann, 1967). 71

91 From the above discussion it is evident that institutions have more than one dimension and could be understood as a process (institutionalisation over time) or an exogenous variable. Scott (1995; 2001) created an analytical framework that encompasses different dimensions of institutions and divided them into three interdependent pillars, namely the normative, cognitive-cultural and regulatory pillars. The following section discusses Scott s (1995; 2001) framework. Institutions: definition and Scott s institutional pillars According to Scott, institutions can be regarded as relatively stable entities which may be "subject to change" by social actors. They consist of three main pillars regulatory, normative, and cognitive-cultural which are present at both micro and macro levels. Each of these pillars signify important aspects such as rules (regulatory), norms (normative) and beliefs (cognitive) which are usually disseminated though different types of "carriers" such as formal, relational, and symbolic systems (Scott, 2001: 48). Furthermore, institutions can be classified as both formal and informal. At one extreme lie formal institutions, which articulate formal rules, constitutions, specific bylaws and contracts defining the terms and boundaries of a specific transaction or exchange between concerned parties (North, 1990; Scott, 1995; 2001). At the other end of the spectrum lie the informal institutions, which are bearers of "codes of conduct, norms of behaviour and conventions" (North, 1990: 36). Informal institutions are often not as obvious and clear as in the case of formal institutions. They emerge as a consequence of the imperfect information that is diffused within society. This gives scope for individuals to interpret information based on their experience and in this way informal institutions tend to reflect the culture of the country in question (North, 1990). Moreover, informal institutions are represented by a set of 72

92 expectations and norms that govern business practices and codes of conduct within particular "localised interpersonal relationships" or groups of actors (Scott, 1995; 2001:48). Scott (1995; 2001) offers an encompassing analytical framework that covers both the formal and informal aspects of institutions. The formal/regulatory institutions which have to do with laws, regulations and sanctions are coercive mechanisms that encourage or discourage certain behaviours (Scott, 1995:2001; Busenitz, Gomez, and Spencer, 2000). By contrast, informal institutions are composed of values, norms, beliefs and symbols that define desirable or expected behaviours, shared and cultural supported beliefs, and the ways of doing things embedded in society, which are passed on from one generation to another (Boyd and Richerson, 1985; North, 1990; Scott, 2001). Informal institutions in Scott s (1995) terms encompass both cultural-cognitive and normative dimensions. He differentiates between culture and normative dimensions, and argues that culture involves the diffusion of knowledge sets, values and norms of behaviour from one generation to another through the process of copying and instructing (Boyd and Richerson, 1985; North, 1990). These shared understandings and norms are the framework through which individuals and organisations make rational decisions and choose a certain course of action. However, normative values and norms are more related to moral obligations that are deemed desirable or undesirable by a particular group of people such as a certain religious groups, members of a professional body. These normative institutions stipulate the appropriate course of action and whether this course of action is likely to be "certified or accredited" by the members of a social group (Scott, 2001:52). 73

93 Scholars from different disciplines have discussed the impact of institutions on the level of innovation within countries, on how international partnerships evolve, and on how entrepreneurial activities differ from one context to another (Kiss and Danis, 2008, Schumpeter, 1926). Within the international entrepreneurship literature, studies have focused on how social networks influences entrepreneurial initiative in countries like China which are characterised by a collectivist culture and low level of institutional development. According to Kiss and Danis (2008) there is still limited understanding about the impact of different national contexts on the internationalisation process of the firm and the extent to which firms rely on their social networks (Jones et al., 2011). Institutions, social capital and the entrepreneurial behaviour of firms Countries differ in their level of institutional development. This disparity has implications for the extent to which entrepreneurial behaviour is encouraged or restricted. In addition, it is well established within the network literature that relationships are context specific and that they are embedded within their local context or environment (Kiss and Danis, 2008; Dodd and Patra, 2002; Jones et al., 2011). This means that the country s level of institutional development is likely to determine the extent to which entrepreneurs rely on social relationships to do business and whether local institutions (formal or informal) encourage or discourage entrepreneurial internationalisation activities. Table 2.6 provides an explanation on how the different levels of institutional development affects the SMEs internationalisation and networking behaviours. 74

94 Table 2.6 Relevance of institutional development for internationalisation and networking Institutional Dimension Definition High Level of institutional development Low Regulatory Refers to governmental policies and regulations that govern the cross border activities of firms (Busenitz et al., 2000). It also identifies other formal institutions that play a crucial role in offering support for firms such as equity markets and banks (Kiss and Danis, 2008; Busenitz et al., 2000). Well established legal and monetary conditions demonstrated in effective law enforcement and efficient banking systems and efficient support provided through government supporting agencies to internationalising firms. Thus creating a favourable business environment (Bruton et al., 2005) It is difficult for entrepreneurs to internationalise with unfavourable regulatory conditions, such as export restrictions and restrictions on finance. Normative This sets the standards and values shared by members of a group or organization regarding the preferable way to do business (Scott 1995; 2001; Kiss and Danis, 2008; Burton, Fried, and Manigart, 2005). Internationalisation is stressed as a desirable behaviour and an important means of business development. Internationalisation is less valued and there are limited normative values supporting internationalisation. Moreover, it also involves the expectations and accreditation of informal professional groups such as members of the same industry. In the context of entrepreneurship, normative institutions, such as export promotion agencies, potentially influence the importance that entrepreneurs place on internationalisation and view it as an integral aspect of their overall business (Kiss and Danis, 2008). Cognitive/ cultural This identifies cultural values and norms that govern the way entrepreneurs generally do business (in terms of honesty, integrity and how to deal with foreigners). It also reflects the importance of relationships in society and more specifically as a means of doing business (Bruton et al., 2005; Kiss and Danis, 2008). Relationships may be important. However, international expansion and firms success is more the primary concern/ in other words, the entrepreneurs would not focus on creating strong relationships first with their business partners before they start talking about business. Hence the focus on weak ties is prevalent (Dodd and Patra, 2002; Dodd, Jack and Andersson, 2002). Personal relationships are more important than doing business. Hence strong ties are crucial for doing business (Dodd and Patra, 2002; Dodd et al., 2002). 75

95 The lower the institutional level of development the more likely it is that firms will rely on strong personal ties to compensate for intuitional voids. Conversely, the higher the level of institutional development, the more they will rely on weak ties (Kiss and Danis, 2008; Dodd, Jack and Andersson, 2002; Dodd and Patra, 2002). However, this raises an important question: if this is the case during the formation stage or early establishment of the firm, will it apply to later stages? The constraints that a country s institutional framework imposes on individual or organisational actions are the sum of both formal and informal institutions. It is important to note that not only do institutions create opportunities and constraints that influence the entrepreneurial firms behaviour over time but also entrepreneurs (political or economic) influences and alters the institutional framework in a way that benefits them. This triggers institutional change that is incremental and stable in nature, whereby laws are revised and altered in a gradual manner. In some instances this process of change could be on a discontinuous (revolutionary) basis whereby, a dramatic change occurs to formal or informal institutions (North, 1990; Scott, 1995; 2001). The main focus of this study is on how the institutions facilitate or hinder entrepreneurial (internationalisation) activities and how entrepreneurs alter and influence the institutional framework in a way that creates value for them. Institutional change and paths through which it evolves is beyond the scope of the current study. The present study focuses on two different national contexts which exhibit contrasting levels of institutional development, namely UK and Egypt. The above discussion suggests that the 76

96 level of institutional development will impact on SME internationalisation and networking behaviour. The following section briefly reviews the contextual differences between the two countries. National institutional differences between the UK and Egypt The emerging literature examining the impact of institutional contexts and national differences on entrepreneurship has been controversial. Some scholars argue that different institutional frameworks and logics have a strong impact. Others often referred to as Universalists adopt a generic view of entrepreneurship which assumes that entrepreneurs from different countries share more behavioural similarities than exhibit national differences (Dodd and Patra, 2002; McGrath and McMacmillan, 1992; McGrath, MacMillan, Yang and Tsai, 1992). The latter group argues that a universality of entrepreneurship behaviour is grounded in the fact that entrepreneurs have some generic characteristics and behaviours which prevail across different contexts. However, several empirical studies have challenged this view, especially ones covering Chinese and Asian entrepreneurs (Kiss and Danis, 2008). Asian entrepreneurs and SMEs are strongly embedded in collectivist relationship-based cultures, where being a member of a certain social network is a primary key to success (Dodd and Patra, 2002; Kiss and Danis, 2008; Hofstede, 1983; 1991). In these contexts "guanxi, or connections with people in senior positions such as governmental officials, play a significant role in facilitating and supporting SMEs' international survival and growth (Kiss and Danis, 2008). 77

97 The former set of scholars argue that differences in institutional frameworks and contexts affect the way firms operate in the home market as well as having a significant consequence for their internationalisation (Boddewyn, 1988; Boddewyn and Brewer, 1994). Scholars have noted that home country institutions, both formal and informal, shape how firms do business in domestic markets. SMEs have to conform to bureaucratic rules and polices laid down by local institutions, as well as relying on their support to internationalise overseas (Buckley, Clegg,Cross, Liu, Voss and Zheng 2007; North 1990). These scholars emphasise the importance of different contexts and institutional frameworks in shaping the nature of SME internationalisation behaviour (Kiss and Danis, 2008; Dodd and Patra, 2002; Shane, Kolvereid and Westhead, 1991). Furthermore, Powell and DiMaggio (1991) argue that institutions within home and overseas markets can hinder firms internationalisation through enforcing laws and ways of doing business that could obstruct their internationalisation. Whereas others argue that the institutions in local and overseas markets are crucial in assisting SMEs internationalisation (Child and Rodrigues, 2005; Whitley, 1999; Buckely, Clegg, Cross, Liu, Voss and Zheng, 2007). Institutional theory and its dimensions The most widely used framework of institutional dimensions is that advanced by Scott (1995). The dimensions he identifies overlap and are not mutually exclusive (Hirsch and Lounsbury, 1997; Bruton, Fried and Manigart, 2005). As discussed above, these are the regulatory, normative and cognitive "pillars" of institutions. Bruton et al. (2005) compared the three dimensions in different regions and/or continents and examined their impact on the expansion of venture capital companies in the US/UK, Asia and continental Europe. They argued that in mature and strong institutional contexts 78

98 entrepreneurship is highly regarded and that firms rely to a lesser extent on social networks. On the other hand, in weak and less mature contexts (such as Asia and most African countries) entrepreneurs occupy a relatively low status and they tend to focus aggressively on social connections to fill in institutional voids inherent in their respective contexts. In such contexts, entrepreneurs and small business owners tend to have strong inter-personal relationships within key networks embedded in their domestic institutions as an alternative route to enact and implement contracts and sometimes even to impose "sanctions on violators" (Bruton et al., 2005: 743; Perkins, 2000). Hence, entrepreneurs are more likely to spend more time nursing and nurturing these key relationships than on building new ties or focusing on short term profits (Chen, 2001; Bruton et al, 2005; Dodd and Patra, 2002). To date, few studies have examined the impact of different institutional conditions on SME internationalisation and networking behaviours (Kiss and Denis, 2008; Dodd and Patra, 2002), and these studies have focused on developed countries and hardly tackled any emerging economies. Emerging economies are more likely to suffer from institutional voids which refers to absent or weak specialized intermediaries, regulatory systems and contractenforcing mechanisms in their product, labour and capital markets (Khanna and Palepu, 1997; Khanna and Palepu, 2010). They often manifest as gaps between formal rules and norms, and their enforcement in daily practice (Rodrigues, 2013: 14). Most attention to institutional voids in emerging markets has been given to the challenges and opportunities they offer to firms seeking to do business in such markets. The effect of domestic institutional voids on internationalising SMEs has received next to no attention. 79

99 According to the Global Entrepreneurship Monitor's extended report (2011), entrepreneurship levels vary significantly between what it terms factor-driven and innovation-driven economies. The former consist of African and Middle Eastern countries such as Egypt and Tunisia, and the latter consists of developed western countries such as the USA and UK. An indication of the level of institutional voids in various emerging and developed economies is provided by the World Justice Project s Rule of Law Index (Agrast et al. 2013; Rodrigues, 2013). Table 2.7 shows the scores for UK and Egypt. 9 Table 2.7: Indicators of the extent of institutional voids in Egypt and the UK Open government Regulatory enforcement Egypt UK These scores indicate that Egypt has considerably greater voids in its institutional system than does the UK. This is consistent with data for 2010 obtained by the World Economic Forum (2011) from its Executive Opinion Survey, which score Egypt considerably higher than the UK for indicators of institutional voids such as corruption, irregular payments and bribes, favouritism in decisions of government officials, and lack of transparency in government policymaking. 10 Accordingly, an important rationale for the design of this study is that different institutional frameworks are likely to have an impact on the international and networking behaviour of 9 Data are for 2012, since earlier information was not available for Egypt. The index is constructed so that 1=the highest possible score and 0=the lowest. Open government is assessed by three indicators reflecting whether laws are publicised, stable and accessible. Regulatory enforcement is assessed by three indicators indicating whether regulations are enforced effectively, without improper influence and unreasonable delay. 10 The reason for using data representing the situation of Egyptian economy in 2010 is that these are the latest scores before the 25 th January 2011 revolution and because this also approximates the time (December 2010) when interviewing was completed in Egyptian SMEs. 80

100 SMEs and those different levels of institutional development are key contextual features. The next section describes in further detail differences between UK and Egyptian institutions, including their cognitive-cultural aspects, and how their anticipated relevance for SME internationalisation and networking. Contextual differences, SME internationalisation and networking As mentioned before, Scott s (1995; 2001) institutional pillars could be grouped broadly speaking into two main categories which hare formal and informal institutional pillars. Where the former refers to the regulatory dimension and the latter refers to normative and cognitivecultural dimensions. Cognitive-cultural informal institutions These aspects of institutions consist of the cultural and normative values that shape entrepreneurs' behaviour, and establish understandings as to what constitutes fair and acceptable business practice. SMEs and entrepreneurs usually rely on networks and contacts to compensate for the scarcity of resources and limited experience in order to survive and grow inside and outside their home country. These relationships could be purely of a business nature such as those with suppliers, customers, and agents and supporting agencies; or they could be social relationships such as those relationships with family, friends and colleagues. These relationships are highly likely to be informed by cognitive understandings and cultural values. Although cognitive-cultural similarities might exist at the national level, such as among so-called developed Anglo-Saxon countries, differences between groups of countries may be expected to shape networking behaviour in addition to variations due to the unique 81

101 background and idiosyncrasies of each individual entrepreneur (Birley, Cromie and Myers, 1991; Dodd and Patra, 2002). Over the last decades, Hofstede s (1983; 1991; 2001) five cultural dimensions have been widely used to explain variations in national cultures. He suggests that each nation or country exhibit at least some different cultural and social characteristics. These cultural dimensions are power distance, uncertainly avoidance, masculinity, Long term orientation and individualism. Individualistic versus collectivist cultures Individualistic societies are loosely tied to groups of individuals, whereby "I" is emphasised and dependence on groups of people is not common. On the other hand, collectivist societies are those that typically have tightly or closely linked groups of people (in-group). Also, being part of this group offers safety and is usually characterised by strong loyalty and trust. Each member of this tightly-knit web of relationships is expected to look after the interest of the whole group rather than his or her own interest. The UK Scored 89 on this dimension, ranking it along with the United States, one of the most individualistic cultures Hofstede (2001) 11 On the other hand, Egypt scored 25 which indicate that it is a collectivist culture. 11 These scores are according to Hofstede G. (2001), Culture s Consequences: Comparing Values, Behaviours, Institutions, and Organizations across Nations. 2 nd ed., Thousand Oaks CA: Sage. 82

102 Egyptian culture has for decades focused on the importance of having close and strong relationships with members of family, friends, acquaintances and members of a particular group or club. Establishing personal long term relationship with key strategic individuals and others who might occupy a senior position in the government (and/or have good connections with gatekeepers in crucial organisations) is considered one of the most important means of doing business. Nurturing and maintaining strong close relationships with members of one s social network is a key characteristic of a collectivist relationship based culture (Hofstede, 2001). Hence, Egyptian entrepreneurs are assumed to have a stronger tendency towards maintaining long term relationships over maximizing short term profits. The entrepreneurs network is more likely to be composed of strong social ties with family members, friends and/or members of a particular group. Alexandria, the second biggest city in Egypt, is well known for its closely knit community, especially between businessmen from the high income class. In this closely knit community everyone knows everyone else and the exchange of tacit and strategic information is common. For example, non-members of the Alexandria Rotary Club are greatly disadvantaged and they do not enjoy the same privileges as members of the club which is mainly composed of high class well educated business men. 12 Other parts of Egypt, Cairo for instance, are characterised by larger and less integrated communities. However, entrepreneurs tend to form ties with well-connected individuals who either know or have a direct relationship with individuals in senior/strategic positions in government, public sector companies and/or key gatekeepers in different organisations. Broadly speaking, Egyptian entrepreneurs rely on strong trust-based relationships they have with family members, 12 Key membership advantages such as helping the company to grow and to obtain necessary licences and registration quicker than its counterparts was emphasized by three entrepreneurs from my Egyptian sample Moreover, the closure effect was reported by another Egyptian entrepreneur interviewee. 83

103 kinship, colleagues and acquaintances in order to obtain support and access to key information and resources that are crucial to business survival and growth. In Egypt the term wasta refers to finding someone in a high position to get things done faster and even to obtain approvals and facilitate the processing of necessary documentation (Hutchings and Weir, 2006). Given the collectivist nature of the Egyptian culture, it expected that entrepreneur networks will be interconnected where alters in an ego s network know one another. Friends of friends know one another; this is typical characteristic of closed (dense) networks. These strongly interconnected dense networks have been found to have a positive impact on SME growth (Hansen, 1995). The type of information accessed and exchanged through these network ties is likely to be tacit in nature and is expected to be transferred quickly between members of a given network. These relationships take a long time to develop and are characterised by high levels of trust and cohesion (Naphabiet and Ghoshal, 1998; Coleman 1988). This is similar to the findings of Dodd and Patra (2002) who found that Greek entrepreneurs (Greece is considered a collectivist culture) tend to invest a lot of time nurturing close and strong webs of relationships in their networks which makes it difficult to find enough time and resources to build new ties. Moreover, they argue that entrepreneurs who are not members of these close-knit networks are disadvantaged by not being able to gain access to important support and information. This is closely linked with Coleman s (1998) closure phenomena, whereby exclusion is a key disadvantage for people who are not connected to such a closed network (Han, 2006). 84

104 Egypt is highly corrupt and suffers from a weak institutional framework, which makes it difficult for Egyptian entrepreneurs to survive. One way to compensate for institutional voids is to rely on bridging ties or brokers (Burt 1992; Estrin, Bhaumik and Peng, 2008). These are usually weak business ties that serve as a bridge for the company to gain access to crucial resources such as key governmental support. By contrast, the UK is considered a highly individualistic culture whereby people tend to be more idiosyncratic and value their privacy over social relationships. In other words, being a member of an "in-group" is not as important as with the case of collectivist cultures (Hofstede, 1991; 2001). It is a norm of British culture that each person should contribute in a unique and individual way to the society. Since the UK has a strong and mature institutional framework (Bruton et al, 2005), British entrepreneurs are expected to focus more on weak contractual ties to gain access to resources and information and to protect against opportunistic behaviour and risks. Transaction related aspects are made explicit in formal contracts and documents. Relationships with different clients, suppliers and so forth are mainly business relationships, whereby personal and work related aspects do not overlap as it is the case in a collectivist culture (Dodd and Patra, 2002). Also entrepreneurs tend to be more competitive and individual success is their primary goal in life. Hence, membership of a certain group or network is not their primary aspiration but rather the short-term profit and growth of their companies (Hofstede, 2001; 1991). 85

105 The entrepreneurs network is more likely to be a loosely knit network composed mainly of weak short term business ties. Individualistic societies tend to place lesser value on maintaining relationships compared to collectivist cultures. This might be loosely linked to the usefulness cycle or utility life cycle concept, whereby once a relationship has fulfilled the purpose for which it was initially established, the relationship ceases to exist (Prashantham and Dhanaraj, 2010). Furthermore, it is less problematic for British entrepreneurs to choose their partners or relationships as there are comparatively fewer institutional voids than in Egypt. They are more likely to choose the tie that will offer the best benefit for them. On the contrary in collectivist cultures, favours and social aspects of relationship tend to be a priority over economic benefits. Finally, British entrepreneurs are expected to have more time to nurture and establish new ties as their networks are mainly composed of weak business ties. Therefore, are expected to have more sparse but large networks compared to collectivist cultures. Power distance High power distance cultures are ones that tend to acknowledge hierarchy and in which members of the society accept that it is appropriate for some individuals to occupy higher status or privileged positions while others do not. In other words, they are more understanding and tolerant towards inequality within society (Hofstede, 1991; 2001). In high power distant countries entrepreneurs might perceive that opportunities and support are available only to certain people. In Asia and the Middle East entrepreneurship is considered something that only the elite of the society undertake, especially owners of big 86

106 firms. This makes average entrepreneurs (leaders of small firms) more reluctant to seize new opportunities as they emerge either domestically or internationally (Mitchell, Smith, Seawright. and Morse, 2000; Dodd and Patra, 2002). By contrast, low power distant cultures tend to de-emphasize status and other hierarchical differences between people, and strong preference is given to minimizing inequalities in the society. Egypt had a score of 70 which makes it a relatively high power distant country (Hofstede, 2001). This is evident in the way entrepreneurs do business and how they approach opportunities, conflicts and problems as they arise in the domestic environment. They tend to think in terms of "who do I know or do any of my friends and/or family know someone who could help me" rather than "what are the formal guidelines I should follow" (Dodd and Patra, 2002). Egyptian entrepreneurs tend to spend more time socialising and nurturing key strong relationships which limits their ability to have large- sized networks. On the other hand, they might need over time to form several weak brokerage ties in order to obtain access to other closely connected networks to which they are not linked to so as to gain access to support, information and other resources not available in their closed networks (Burt, 1998; Han, 2006; Dodd and Patra, 2002). By contrast, the UK scored 35 which makes it a low-power distant country, where inequality in the society is minimized and hierarchy is so not dominant (Hofstede, 2001). This suggests that British entrepreneurs are less likely to form relationships with partners who are well 87

107 connected or occupy a prominent position in the society. In other words, it does not matter which position their partners occupy; what matters is the business at hand. 13 Uncertainty Avoidance Uncertainty avoidance has to do with the extent to which members of a particular society are risk takers or risk averse in terms of dealing with the uncertain future. Some people feel at ease with uncertainty and ambiguity. People in countries typified by a high level of uncertainty avoidance tend to avoid change and innovation and prefer to be guided by clear rules and guidelines; although these rules may not necessarily be fully enacted (Hofstede, 1991; 2001) Egypt scored 80 which indicate that Egyptians tend to be risk averse individuals and prefer to avoid ambiguity (Hofstede, 2001). Egypt suffers from continuous change in its laws and regulations as well high level of corruption. This is a strong indication of a weak institutional framework. To compensate for this, Egyptian entrepreneurs are more likely prefer to establish strong long term relationships characterised by a high degree of trust with alters in their network. These alters are more likely to have had a history of positive interaction with entrepreneurs, which makes it easier for the entrepreneur to assess whom to get in touch with when a particular problem arises or an opportunity or favour is warranted (Dodd and Patra,2002). 13 This point was mentioned in the second pilot interview, which was in a British oil and gas company. 88

108 Given the personalised nature of these ties the entrepreneur will find it easier to seek resources, exchange favours and seek tacit and sensitive information than is the case with weak ties. However, in order to keep these relationships enacted and maintained, the entrepreneur is more likely to spend more time nurturing these relationships than with weaker ties. Moreover, they tend to become more rigid in terms of changing their existing social networks (Dodd and Patra, 2002). Furthermore, most Egyptians are known to be risk averse by nature and tend to prefer to deal with the familiar and to maintain the status quo. Egyptian culture is a fatalistic one (Elbanna and Child, 2007). Entrepreneurs would prefer to choose opportunities that involve lowmedium risk and to operate in similar environments. 14 Hence, Egyptian SMEs are expected to internationalise to nearby markets or to ones with which there is a low psychic distance (such as other Middle Eastern countries). Similar environments or countries with a lowpsychic distance have some commonalities in consumer tastes, language, laws and so forth. In addition, Egyptian SMEs are expected to depend on direct or indirect exporting rather than having overseas agents and/or wholly owned subsidiaries. This is attributed to their preference to keep risk to a minimum. Alternatively, they may resort to personal contact(s) or relationship with Egyptian immigrants in overseas markets as an alternative way to enter new markets. By contrast, the UK scored 35 which is a low uncertainty avoidance score. This implies that British entrepreneurs tend to be risk takers and more tolerant of uncertainty and ambiguity (Hofstede, 2001). Accordingly, one would expect that they tend to be more innovative and 14 This is a general tendency and there are exceptions to it. Some Egyptian entrepreneurs are opportunistic and risk takers. 89

109 accommodating to change. They are willing to take on board risky opportunities when it comes to internationalising into overseas markets. They are more likely to internationalise to a diverse number of markets and in markets where psychic distance is high. Moreover, the UK institutional framework is more mature and stronger than that of Egypt and this provides more formalized ways of avoiding uncertainties and opportunistic behaviour on part of their partners, such as soundly-based contracts. This is enhanced by the individualistic nature of British entrepreneurs whereby they are more likely to rely more on weak business ties of shorter duration (Dodd and Patra, 2002). In other words, the utility life cycle of some of the ties is short (Prathanham and Dhanaraj, 2010). For example, if an entrepreneur has a relationship with a manager of a MNC and both companies are working on a medium term project, once this project is finished and there are no further opportunities coming out from this relationship, the tie with the MNC manager is more likely to decay as its usefulness has finished. On the contrary, if an Egyptian entrepreneur has a close personal relationship with the project manager, the relationship is more likely to continue even after the project finishes. Masculinity This dimension has to do with drive to excel at work and being success driven. Being more of a feminine culture, means that there is a stronger need for support and compassion and enjoyment from being a member of the overall group (Hofstede, 2001). The UK had a score of 66 which makes it a relatively masculine culture. British entrepreneurs are expected to have a strong drive for success (Hofstede, 2001). Hence striving to win or be highly successful is a key characteristic of British entrepreneurs. This is closely linked to the individualistic nature of British culture and its risk taking propensity, whereby entrepreneurs 90

110 are more likely to focus more on business and on creating business relationships that yield the best benefit(s) and provide access to key resources than are their Egyptian counterparts. Egypt has a score of 45 which makes it a relatively feminine culture (Hofstede, 2001). In Hofstede s terms it is a culture that values support, solidarity and relationship quality. Moreover, conflicts are solved via reaching compromising solutions and through diplomatic efforts rather than winning over. This implies that Egyptian entrepreneurs are more likely to value the social aspect of their relationships and seek supportive ties rather than merely business relationships. This is closely coupled with the collectivist nature of Egyptian culture and its risk avoidance tendency, whereby trustworthy, strong personal relationships become the key characteristics of Egyptian entrepreneurs networks. Long- term orientation This dimension has to do with certain cultures being more focused on long-term goals and the future rather than short term goals. Also, preference is given to conventional traditions and history (Hofstede, 2001). The UK scored 25 on long-term orientation, indicating that it has a short term oriented culture focusing on short to medium term goals. However, Hofstede did not provide any scores on this dimension for Egypt or any other Arab country. Egyptian entrepreneurs are likely to be fatalistic and believe that the future is based on one s destiny (Elbanna and Child, 2007; Trompenaars, 1973). This is reflected in the way they view and seize opportunities as they arise. Egyptian entrepreneurs are more likely to be slow in deciding the future course of action. This is more likely to result in a gradual internationalisation into new overseas markets. 91

111 However, one would expect British entrepreneurs tend to focus on achieving short term goals and to be more proactive in creating opportunities (Hofstede, 2001). They are more task and short term oriented than Egyptian entrepreneurs. The sense of achievement and being the best is a key cultural driver. This is more likely result in leapfrogging or accelerated internationalisation of the firm. Hofstede s five cultural dimensions are interrelated and influence one another. The scores ascribed to Egypt and UK is summarised in Figure UK Egypt PDI IDV MAS UAI LTO 0 *Adapted from (Hofstede, 2001). Figure 2.4: Scores for Egypt and the UK on Hofstede s Cultural Dimensions To summarize: in view of their national cognitive-cultural characteristics, Egyptian entrepreneurs are more likely than those in the UK to have a dense network composed of strong and close long term ties. It is expected that they draw heavily on friends, family, kinship, colleagues and acquaintances to gain access to knowledge and resources and to solve problems that they encounter as a result of the institutional voids inherent in the Egyptian 92

112 institutions. If they cannot find the solution or what they are looking for in their tightly knit network, they would rely on brokers or referrals (usually through friends or family members) to form weak ties with individuals in key positions and/or who are key gatekeepers to access the needed support and resources. Maintaining and nursing these strong relationships is expected to constrain the entrepreneur from forming new ties (Dodd and Patra, 2007; Burt, 1998; Coviello, 2006). Egyptian entrepreneurs are also likely to be risk averse and fatalistic. They would prefer to take gradual reactive steps when it comes to dealing with risks associated with internationalisation. They are more likely to expand into similar markets which are characterised by low psychic distance and who are culturally similar to Egypt. Moreover, they are expected to opt for the least risky mode of operation in overseas markets and/or to depend on their strong network ties to enter foreign markets. Some may draw upon the immigrant population in overseas markets that are largely dissimilar to the Egyptian context. One key disadvantage of Egyptian entrepreneurial networks is that they are composed largely of people who share same attitudes, information sources and ideas. This creates redundancy in their main information sources and might result in the closure effect highlighted by Coleman (1988). They are less likely to have access to new sources of information and opportunities (Burt, 1992). By contrast, UK entrepreneurs are more likely to have sparse, non-cohesive networks composed mainly of weak business ties. They will draw heavily on contracts and other means of hedging themselves against the risks associated with internationalisation. They are more 93

113 likely to be less constrained than their Egyptian counterparts and have more time to build and establish new ties and have access to diverse sources of information and resources. They are more proactive and risk taking, responding quickly to opportunities as they arise in overseas markets. This makes the overall entrepreneurial network less rigid and open to new sources of information, either through proactively creating new weak ties or through relying on brokers to fill in the structural holes inherent in their network (Burt, 1992). Normative informal institutions Normative values and shared patterns of behaviour within a particular context influence and shape the behaviour of organisations and individuals (Scott, 1995a; March, 1981). SMEs embedded in a particular context are importantly shaped by normative institutions in the form of professional or governmental institutions and financial institutions. These institutions can help promote entrepreneurship or hinder it. This is closely related to the way a particular society values entrepreneurship. According to the Global Entrepreneurship Monitor report (2010), the total entrepreneurial activity (TEA) rate in Egypt was 7.01%. 15 However, the TEA rate in the UK was 12.7%. This indicates that there is a higher percentage of entrepreneurship in the UK than there is in Egypt. Nevertheless, Egyptians strongly believe that they can succeed and run a small entrepreneurial business. This is an inherent norm among recent graduates or young adults who think it is better to start their own business and be their own boss rather than working for 15 Total entrepreneurship activity (TEA) is the "percentage of total adult population (aged between years) who have either recently started a business, have been running a business (for three and half years) or planning to run a business" (GEM Egypt Report, 2010:6). 94

114 someone who might control him or her. 16 Young adults strongly believe that they will succeed and their business continue to grow, but they also think that the society and particularly professional bodies -such as local chambers of commerce -will not support them and might even impose restrictions on them. Many Egyptians want to be entrepreneurs but the institutional environment does not offer enough normative support to foster entrepreneurship. In the UK, entrepreneurs are highly regarded and many young adults have the drive and determination to try new innovative ideas and are proactive in creating opportunities (Dodd and Patra, 2002). The financial support available to SMEs whether from public or private sources varies in both countries. According to the GEM score for the degree of financial support offered to SMEs, Egypt was ranked 27 th and UK was ranked the 34 th, which places both countries in the middle range of the 59 countries examined in the GEM (2010) report. It suggests that the financial support offered to SMEs is relatively inadequate in both countries, although venture capital and private equity companies are not as common in Egypt as they are in the UK (GEM Egypt report, 2010). In general, overall social norms and values might hinder or foster entrepreneurship (GEM Egypt report, 2010). Egypt occupied 51 st position (alongside Turkey which had the lowest score) making it one of the most constraining environments for entrepreneurship. By contrast 16 Based on interviews conducted in 2006 with some recent graduates from the Faculty of Commerce English section, Cairo university the researcher s former employer as a part of a project on young adults career prospects. 95

115 the UK occupied 19 th position which is a relatively high score, indicating an encouraging environment for entrepreneurship. This type of informal institutional context is difficult to investigate and to separate from cultural-cognitive dimensions. For the sake of simplicity in this research I will use informal institutions in their broader sense and focus on how entrepreneurs perceive the institutional environment as favourable or unfavourable. It is argued that the abundance of institutional voids and the ways SME decision makers cope with them become institutionalised over time as part of the culture and norms of the society. Formal institutions Formal institutions refer primarily to government policies, laws and property rights that have an impact on SME behaviour. Government policies and provision to support SMEs could also be regarded as part of the formal institutional field (GEM Egypt report, 2010). The efficiency of government support programmes to help and offer support to SMEs contrasts between the two countries. According to the GEM (2010) report Egypt occupied 41st position in terms of the efficiency of the government support programmes which indicates a rather poor effectiveness of these support programmes. On the other hand, the UK occupied 20 th position which is an indication of relatively strong governmental support programmes. In the UK over the last decade there has been a growing interest in supporting SME development and growth. Part of the government initiative is to help and support British SMEs to export and internationalise through the UK Trade and Investment teams [UKTI] and their close links with the network of local chambers of commerce. The UKTI organization is that part of British government focusing on offering help and support to UK based SMEs to internationalise in overseas markets. It has a network of international trade advisors who 96

116 work closely with local chambers of commerce across the UK and also have links with overseas embassies/consulates and high commissions. Several programmes and opportunities are available to exporting SMEs and some are targeted to experienced exporters. Examples of key programmes and support offered to the UK SMEs are summarized in Table 2.8. Table 2.8: Examples of government programmes and support offered to UK SMEs Services /Programmes Details /comments Targeted at UKTI general advice from international trade advisors Passport to Export programme Gateway to global growth Overseas market introduction scheme Help SMEs to have access to information, advice from UK based staff and in overseas consulates, help in identifying business leads in overseas markets, sources of funding available for SMEs to participate in trade fairs and other financial assistance programmes. A yearly programme to help companies to create their unique export strategy; cultural awareness and documentation. It also offers advice on assessing the demand for a company s product in overseas markets A kind of strategic review of the firm s export strategy and performance. It offers more advanced advice on enhancing overseas sales growth, training and information about different modes of operation and what each of them entails. A service that offers help to exporters to get in touch with overseas consulates and embassies; market knowledge and links with key potential buyers; and financial support that helps SMEs to participate in overseas trade missions and specialised fairs. Mainly new exporters but some advice is also given to established and experienced exporters. New exporters Experienced exporters Both new and experienced exporters Source: UKTI and West Midlands Chamber of Commerce brochures and website (2010) Table 2.8 shows that a diverse range of services and support programmes is offered to SMEs and points to a strong regulatory framework that helps and fosters entrepreneurship. During the last decade (up to the Egyptian revolution of early 2011), the Egyptian government has been increasingly interested in promoting SMEs. Law (no.141/2004) was introduced and focused on developing small enterprises. It states that the Social Fund for Development (SFD) is a primary source of governmental support for SMEs. The SFD has the 97

117 legal responsibility to network with local chambers of commerce to offer support for SMEs across Egypt. The SFD offers some help to SMEs in general in terms of giving advice on successful projects; offering assistance in preparing feasibility studies; financial support; and referrals of its beneficiaries to key governmental banks. These banks offer the SFD s beneficiaries subsidized loans. Moreover, SMEs which are eligible for financial support receive a five year tax exemption and the right to become partially sponsored to participate in the trade fairs organised by the SFD. Exporting SMEs get some support in terms of partial sponsorship towards the cost of participating in local trade fairs to which local and overseas clients are invited There is also partial sponsorship towards the cost of participation in overseas Egyptian product fairs; these are non-specialist fairs that display a broad range of Egyptian products in a particular overseas market. Other governmental supporting agencies are closely linked to the Ministry of Investment namely Expo Link and IMC. These agencies are different from the SFD in that they are sponsored by overseas sources of aid such as Japanese or European aid, and usually they have both local and overseas management. They target some sectors and industries, and they offer financial assistance towards the cost of international certifications, overseas trade fairs, and so forth. Thus Egyptian governmental policies and regulatory environment offer some help to SMEs. They are still inefficient but have some potential (GEM Egypt Report, 2010). To highlight key differences between the UK and Egyptian institutional frameworks, some of the GEM entrepreneurial framework conditions (EFC) are presented in Table

118 Table 2.9: GEM entrepreneurial framework conditions (EFC) Entrepreneurial support UK Egypt Comment Availability of financial support Ranked 34 th Ranked 27 th Both are within the middle range of 59 countries examined in the GEM (2010) report, suggesting that the finical support offered to SMEs is relatively inadequate Effectiveness of government entrepreneurship programmes Occupied the 20 th position which is an indication of a relatively effective governmental support programme Occupied 41 st position in which is a poor indicator of the efficiency of these government support programmes. Significant difference between Egypt and the UK. R&D transfer* Occupies 18 th position which is an indication of ease of R&D transfer and collaboration between national research centres and SMEs Occupies 50 th position which indicates poor collaboration between national research centres and SMEs Significant difference Social and Cultural Norms Occupies 19 th position which is a relatively high score, implying an encouraging environment for entrepreneurship. Occupies 51th position alongside Turkey (which had the lowest score) indicating the most unfavourable environments for entrepreneurship Significant differences Source: Adapted from GEM Egypt report, 2010) * R & D transfer measures "the extent to which national institutions and research centres transfer easily R&D to SMEs and help them create new opportunities" (GEM Egypt Report, 2010: 43). Although this measure focuses on new and growing SMEs it gives a general indication of the relative support offered by national research centres and universities. Based on the above discussion it is evident that there are significant differences between the UK and Egyptian institutional context. Both countries have a different institutional logic that is likely to have a strong impact on the SME s behaviour in terms of both internationalisation and networking. The next section discusses co-evolution which is the guiding framework for this research. 99

119 2.4 The co-evolutionary perspective The idea of co-evolution can be traced back to the work of Charles Darwin (1859). Darwin argued that members of a particular species face pressures from their environment which challenge their existence, and that only those species possessing characteristics best fitting that environment survive. Through the process of natural selection, only those species posing the best survival genes are likely to be retained and others will cease to exist. These survival genes are passed on to the next generations which becomes part of their genetic inheritance that is hard to change or adapt to new environmental pressures, creating what Darwin refers to as inertia. The whole population of species evolves as new species emerge creating a variation in the features of these species Darwinism stresses that members of a particular species are passive and have no influence over the host environment (Darwin, 1859; Lewin and Volberda, 1999; Child, Tse and Rodrigues, 2013). Other biologists such as Baldwin (1896) advanced Darwin's evolutionary theory, arguing that species learn through deliberately adjusting to the changes in the environment as they emerge and through finding new means of adaptation that are not necessary inherited from previous generations. This adds intentionality to the classic evolutionary theory and challenges the idea that organisms are passive in adapting to changes in environment. Instead, it addresses the possibility of intentional learning through finding ways to fit and adjust to environment. This highlights two contrasting views regarding the relation between organisms and their environments. The first proposes that the environment is the key determinant of which organisms are likely to survive through the process of natural selection, whereby organisms are passive and have no role in responding to external pressures imposed on them. The 100

120 second view proposes that organisms respond to the external pressures imposed on them, learn for their past experiences and find new ways to adapt to environment (Child et al., 2013; Rodrigues and Child, 2008). This debate among biologists is also present among management scholars who use evolutionary theory to explain how forms of organisations change in accord with changes in the external environment. Organisation selection-adaptation studies, evolutionary economics, institutional theory and population ecology adopt this perspective when attempting to explain how firms, industries and societies response to changes in different environments be these institutional or international environments (Volberda and Lewin, 2003). Scholars from these areas focus on different levels of analysis ranging from macro, mesco and micro levels, pertaining respectively to societies, industries, institutional contexts and individual firms. Co-evolution of firms and external environments Adherents of Darwinism believe in the survival of the fittest. They assume that the environment is the key determinant of the firm s survival and that firms play a passive role in response to external pressures. Furthermore, others argue that firms belonging to a particular sector, industry or institutional context face pressures to which they have to conform otherwise they will cease to exist. These scholars ignore the role of decision maker s strategic choice in finding the best way(s) to adapt to the environment and overlook the possibility of learning from past experiences and findings new ways to adapt to future challenges or pressures as they emerge. This creates a form of "adaptive learning" whereby firms learn how to adapt and find ways to fit their 101

121 organisations to the changes in external environment, through leveraging their strengths to make use of opportunities as they are emerge from environment (Child et al., 2013: 6). Firms may also find ways to overcome threats arising from the environment. More recently, scholars such as Child (1997),Lewin and Volberda (1999) and Volberda and Lewin (2003) view the Darwinian and strategic choice perspectives as complementary rather than competing explanations of how firms co-evolve with their environment. The former focuses more on the macro level (on whole population) addressing the pressures imposed by the external environment (such as sector or country) on members of a particular group. The latter is mainly concerned with the firm (micro) level at which adaptation to external stimuli is a strategic choice made by individual decision makers. Child (1997) argues that strategic choice involves both reactive and proactive actions on the part of a firm's decision makers as to how to overcome constrains imposed on them by the environment. This suggests that both firms and their environments are interrelated and co-evolve together (Murmann, 2010; Lewin and Volberda, 1999). Firms are influenced by pressures from external environment that threaten their firm s survival, while at the same time decision makers reactively or proactively seek to finds ways to adapt to the environments and in some cases they shape the external environment to maximise their benefits. Child and Rodrigues (2011) have argued that how decision-makers in firms can cope with complexities in their environments depends on their power, with small firms having as a result of their limited power to rely on outside support ("mediation"). Hence the importance to them of networking. 102

122 Co-evolution could take several modes, such as a naïve reactive and intentional adaptation mode. The naïve" or purely reactive mode, refers to the case where the environment is the key determinant of which of firms survival (Lewin and Volberoda, 2003:2116; Child et al., 2013). This mode is similar to the prefect competition concept; where it is mainly the competitive forces that decide which firms would survive and which ones will cease to exist (Volberda and Lewin, 2003; Baum and Singh, 1994). The key premise of this mode is that manager do not intentionally adapt or change their companies to the changes imposed on them from external environment. This presupposes blindness on part of the managers to any learning and opportunity to differentiate their firms from others (Volberda and Lewin, 2003). Another mode involves intentional adaption on part of the firm s managers, who focus on adapting to the changes in external environment as these emerge. This perspective emphasises the role of managerial intentionality, where managers of firms follow a conscious cycle of "variation, selection and retention" (Volberda and Lewin, 2003:2118; Campbell, 1994). This means that manager intentionally tend to differentiate themselves from their competitors through continuously adapting, adjusting their firms to changes as they emerge in their external environment. They also tend to learn from past experiences and these experiences tend to shape the way they behave in the future. This mode of co-evolution could, however, still involve a largely reactive choice on the part of managers in response to external stimuli. Child et al. (2013) highlight a third co-evolutionary mode which emphasises the proactive adaptation-manipulation of the environment by the decision-maker. This environmental manipulation involves anticipation of future changes in the environment and choosing the 103

123 best approach to adapt and even change the environment to maximise the firm s benefits and overcome threats and risks before the emerge. Co-evolution of the firm and the environment could be explained by the firm s response to external changes through the processes of exploitation and exploration. Exploration involves searching for new opportunities offered by the creation of new knowledge through innovation, whereas exploitation refers to the use of existing knowledge to enhance the existing processes, products and capabilities (March, 1991). Levinthal and March (1993) argue that to survive companies they need adopt a balance of exploration and exploitation, where the former helps the company to find new opportunities that are vital for their future growth and the latter constructively utilizes existing knowledge to be able to survive at the present time. This means that companies need to avoid either being blinkered by past experiences and processes arising from structural inertia or focusing primary on findings new opportunities and dismissing the importance of learning from past experiences (Volberda and Lewin, 2003). Hence, the need for a balance exploration and exploitation approach whereby decision makers take into account previous experiences and reproduce them in future courses of actions. Also, decision makers should continually engage in finding new ways and opportunities to ensure the future growth and survival of their firms (Volberda and Lewin, 2003). 104

124 It is important to note that co-evolution is different from the organisation adaption-selection theory in that it does not just imply that firms adapt to the environment so that it is the environment that shapes the firm behaviour, but rather the firm could also shape and change the environment through leveraging its capabilities and resources to its advantage. For example, Rodrigues and Child (2003) showed that firms can influence the institutional context in this way. At the heart of the co-evolution concept is the idea that two phenomena interact and influence each other s evolution and development. This interdependency and the interaction do not have to be proportional (Lewin and Volberda, 1999; Murmann, 2003; Pajunen and Maunula, 2008). Thus, in the main, the way the environmental changes influences the firm s evolution is more significant than the firm s impact on environmental change. According to Murmann (2003) and Pajunen and Maunula (2008:249), the primary purpose of the co-evolutionary perspective is to spot the outcomes of the "co-evolutionary interaction between processes" (Volberda and Lewin, 2003). In other words, as Lewin, Long and Carroll (1999) stated, "organizations and their external environments and populations are interrelated outcomes of the firms' managerial actions, institutional influences and extra-institutional changes (such as technological, socio-political and other environmental phenomena" (p. 535). The coevolution perspective tends to focus on how the business firms and their environments evolve in relation to one another. 105

125 Co-evolution and SMEs internationalisation and networking behaviours There is only a limited number of studies applying the co-evolutionary perspective to the evolution of firms and their environment and some are mainly conceptual in nature (e.g. Lewin and Volberda, 1999; Zettinig and Benson-Rea, 2008; Child and Rodrigues, 2011). A common theme among the few empirical studies is the focus on a single case study or industry. For example, Rodrigues and Child (2008) focused on a large Brazilian company and studied how it co-evolved with its political and institutional environment over a long period of time. There appear to be just as few studies using the co-evolutionary perspective to explain SME internationalisation. Exceptions such as Pajunen and Maunula (2008) used a co-evolutionary perspective to study the evolution of new ventures with their industry and external environment. Jones and Coviello (2005) offered a conceptual model integrating SME coevolution with external environment over time based on a set of critical events that SMEs respond to through leveraging their resources and capabilities. In the study of SME internationalisation, the co-evolution perspective offers a "balanced" way to consider the combined outcomes of environmental selection pressures and how managers respond to such pressures through relying on network social capital or changing their internationalisation template (Pajunen and Maunula, 2008:248). It is a framework that crosses different levels of analysis and encompasses the use of different theories with the aim of understanding how firms co-evolve and adapt to the changes in the environment in which it operates over time; whether these changes be at the industry level, or institutional or country level (Lewin and Volberda, 1999; Lewin et al., 1999; Rodrigues and Child, 2008). It serves an integrative framework that takes into account both internal 106

126 processes of the firm and the key drivers/forces inherent in its environment, focusing on the way they interact over time. The main premise of the present research is that SMEs have limited resources and capabilities and tend to rely on network social capital to access these resources that are crucial for its growth and survival. SMEs find it difficult to cope with constraints and challenges imposed on them by the external environment. These constraints and challenges could be inherent in the domestic institutional context and SMEs need to find ways to adapt to these pressures through leveraging their resources and capabilities in a way that accommodates to such challenges and at the same time conforms to the regulatory and institutional pressures imposed on them. One of the key platforms through which SMEs access resources is their network social capital. This is embodied in the relationships that serving as the main conduit through which SMEs can access the resources and knowledge they require, but lack. One way through which SMEs can grow and ensure survival is to diversify to overseas markets, which brings with it internationalisation-related costs and risks. SMEs need to find ways to minimise these risks through leveraging their resources and capabilities. Network social capital can provide one lever available to SMEs seeking to enter new markets and access key resources that are crucial for the firm s internationalisation. SMEs may also be able to reduce the risks in their international business by changing the foreign locations they currently supply through looking for opportunities to enter new overseas markets. This suggests that the way through which SMEs adapt and manipulate the external environment is contingent upon two key processes. The first is the internationalisation 107

127 process which is concerned mainly with SMEs' international operations through directly operating in overseas markets or dealing with an overseas client in the domestic market. The second process is network development which is concerned with the network relationships that SMEs create or re-activate in order to access resources and support to grow domestically and internationally. I have already argued that these two processes are inter-related and influence one another at different points in time. The present research develops a new application of the coevolutionary perspective, whereby firms (SMEs) use and adapt two internal processes to the conditions imposed on them from their external environment. Figure 2.5 illustrates the logic of the co-evolutionary perspective as a guiding framework to explain the study the development of SME internationalisation and networking over time. 108

128 Favourable condition Firm s internal processes External Environment Internationalisation process Shaping actions Institutional context (Home country) Networking process Adaptation actions Institutional and extra-institutional context (Host country) Co-evolution line between firm & environment Unfavourable Condition Co-evolution line between two internal processes Unfavourable environmental condition Favourable environmental condition Firm s proposed course of action Figure 2.5 the co-evolutionary perspective as a framework to study the development of SME internationalisation and networking over time 109

129 While the above framework depicts key linkages between internationalisation, networking and the wider institutional context, the co-evolutionary perspective by definition requires a focus on how these linkages play out over a period of time. Summary and Conclusion The approach taken in this research is eclectic, in that it draws on four different streams of literature as discussed in this chapter. Based on the review of the International Entrepreneurship [IE] and the SME internationalisation literature, little is known as to what happens after the initial internationalisation attempt. Moreover, there is a limited understanding and explanation as to why an SME follows a particular internationalisation pattern and why this might change over time. Some of the internationalisation models highlighted in the first section in this chapter suggests that there are two types of SMEs namely, traditional and international new ventures. The former are small firms that focus initially on the domestic market and as their level of experiential learning increases they start to internationalise into similar overseas markets with low degree of psychic distance and gradually expand into more distant overseas markets. The latter are born global companies that internationalise into a large number of overseas markets immediately or soon after their inception. These two types of SMEs differ in terms of their internationalisation speed and focus. Empirically, scholars have attempted to understand the internationalisation process of the firm through focusing on: (1) the number of overseas markets in which the SME operates; (2) the percentage of foreign sales to total sales and/or (3) a composite measure reflecting the degree of internationalisation. However, internationalisation is multidimensional and 110

130 focusing on one measure per se or on a composite measure could result in over simplification of the of internationalisation process (Kuivalainen et al, 2012). Also, there is increasing evidence of the positive role that the social capital offered by attachment to networks plays in facilitating the internationalisation of SMEs. Network social capital serves as a key platform for small firms to access resources and competencies they lack. However, the ways in which SMEs obtain benefits from their network social capital is under researched. Based on the review of studies that were particularly concerned with the role of networks in facilitating the internationalisation of SMEs, it became evident that these studies focused on: 1. The structural dimension of networks, covering aspects such as the number of contacts in a SME network. 2. General network characteristics such as type of ties, functions offered by different types of relationships, role of weak and strong ties. 3. To a much lesser extent, the quality of relationships, in terms of benefits and costs associated with different types of networks. In addition, the networking process that an SME might follow and how an SME s network change over time remains unclear in this literature. Most IE studies have treated networks either as a dependent or an independent variable, whereby either networks are regarded as the key facilitator of SME internationalisation or networks are shaped by the internationalisation choices of SME decision makers. The ossibility that network and internationalisation co-influence one another has not been considered. For example, at one point an SME might decide to enter a certain overseas 111

131 market and choose an agent from a list of potential agents to help set up a representation office in this market or an overseas client might come across the company s website and place an order. These two scenarios could occur simultaneously at a given point in time. Therefore, SME internationalisation and networking behaviours are likely to co-evolve over time. They do not necessarily follow a uniform or linear co-evolutionary pattern but rather they may move in a cyclical pattern whereby the presence of a key contact in a given overseas market could be the primarily reason why SME decides to enter this market and at the same time the SME might decide to enter another market and search for a good contact in that market to help establish its operations. Previous research has indicated that institutional and cultural contexts significantly influence firm behaviour, because they set the rules of the game and define appropriate and acceptable behaviour. In other words, the institutional context facilitates or constraints the firm s behaviour. Yet, little is known about how SMEs from different institutional contexts rely on their network attachments to cope with the challenges and opportunities presented by the external environment. Based on the review of these three streams of literature, it became clear to me that a comprehensive and holistic understanding of the SME internationalisation and networking behaviours over time is warranted. In addition, understanding how an SME responds to changes in external environment and how the SME might influence the environment for its own advantage is crucial. The co-evolutionary perspective served as an important guiding framework. It allows the researcher to focus on different levels of analysis and integrate different perspectives and 112

132 theories. Empirical studies that have used co-evolution are relatively rare. They have focused mainly on understanding how the firm responds to changes in external environment and to a lesser extent on how the firm influences the environment to its advantage. However, to the best of my knowledge using co-evolutionary perspective to explain how two processes within the scope of the firm s decision-making co-evolve together over time would be novel. To conclude, this comprehensive review of literature provides a general understanding of the variables and measurements used in this study. Moreover, it helps to identify key gaps in the literature and debates such as whether to treat network as a dependant or independent variable. However, to the best of my knowledge, the extant literature does not explain the coevolutionary trajectories of the SME internationalisation and networking behaviours. Also, it offers only a limited understanding of how SMEs respond to changes in external environment (institutional and cultural contexts) and how SME decision-makers might even influence the environment to their advantage. With the above considerations in mind, I deduced two co-evolutionary frameworks for two types of SMEs namely, traditional SMEs and new ventures. These frameworks, presented in the following chapter, are a priori and generic in nature. I propose that traditional SMEs and new ventures follow different internationalisation and network patterns over time and that the institutional context and external environment are expected to shape SME behaviour over time. These are generic rather than precise models and serve as guiding frameworks for data collection and analysis. Based on empirical findings, I shall aim to formulate a revised and more precise framework that provides a comprehensive understanding of the co-evolutionary trajectories of SMEs over time. 113

133 CHAPTER 3 PROPOSED THEORETICAL FRAMEWORK Introduction In an attempt to integrate the temporal dimension, behavioural processes and internal resources Jones and Coviello (2005) formulated a general model to explain the internationalisation of the firm. Following a similar logic, this chapter develops a framework for the present study which focuses on two firm-level processes (internationalisation and networking) that co-develop together and are also seen to co-evolve with the external environment. Two theoretically driven models describing the co-evolution of entrepreneurial firms are developed for this purpose. Each considers co-evolution over three stages of SME internationalisation. The models are a priori in nature having been derived from the existing literature reviewed in Chapter 2. These models are in effect consolidated representations of the propositions that guide the present research. The two models are: 1. The new venture co-evolution model (see Figure 3.1 below). 2. The traditional SME co-evolution model (see Figure 3.2 below). 114

134 Tn Context (institutional and cultural effects) Initial network social capital Evolving network social capital -Structural and relational development Firm Performance Exploration Exploration Exploitation -Learning Preinternationalisation Early Internationalization - No. of countries - Multiple modes - % of foreign sales Subsequent Internationalization 0 T1 T2 T3 Time E1 E2 E3 Figure 3.1. New venture co-evolutionary model 115

135 Tn Context (institutional and cultural effects) Initial network social capital Evolving network social capital -Structural and relational development Firm Performance Exploration Exploitation Exploration &Exploitation -Learning Start-up stage Domestic expansion International Expansion -No. of Countries -% of foreign sales -Multiple modes 0 T1 T2 T3 Time E1 E2 E3 Figure 3.2. Traditional SME co-evolutionary model 116

136 New venture co-evolutionary model When founders decide to start their own business, this may be due to a change in their career or a result of business opportunity that arises for them to capitalise on a unique business idea or know-how they possess. This opportunity could be identified through a friend or an excolleague who suggests the idea to the entrepreneur and then becomes the co-founder of the NV. This trigger point or trigger event is denoted in Figure 3.1 as T1 and E1, where T1 refers to time one and E1 refers to critical event one. The internationalisation process of an NV is assumed to comprise three main stages; the pre-internationalisation stage [PIS]; the early internationalisation stage [EIS], and the subsequent internationalisation stage [SIS]. The model proposes that entrepreneurs will rely on their social capital differently at each stage. The pre-internationalisation stage [PIS] The pre-internationalisation stage of the NV is where the concept or business idea is formed, and the firm is legally established. This is similar to Hite and Hesterly s (2001) start up stage. The entrepreneur begins operations with the strategic goal of how to ensure firm s survival and growth. The firm would explore the existing opportunities within local market and any opportunities to grow outside the home market, often at the same time (Winter and Szulanski, 2001; Barkema and Drogendijk, 2007; McGrath, 2001). Empirical studies have showed that NV founders usually have the capability to scan international markets for opportunities and have an international orientation from the start of their business (Ellis, 2010). However, such firms at this pre-internationalisation stage suffer from limited resources (such as lack of market knowledge, limited financial and non-financial resources) and they also suffer from a liability of newness (Harris and Wheeler, 2005; Chetty and Agndal, 2007; Coviello and Murno, 1996; 1997; Knight and Cavusgil, 1996; Rugman and Verbeke, 2007; Ojala, 2009). 117

137 Thus the firm will start looking for other ways to secure access to needed resources. This is done through the assessment of the entrepreneur s initial stock of network social capital. Initial network social capital The entrepreneur s initial set of relationships consists of the pre-existing relationships of entrepreneur and that of co-founder, which may enable existing social capital to be redeployed (Prashantham and Dhanaraj, 2010). S/he may resort to reusing or redeploying preexisting relationship into a new context. For example, a previous client or colleague who used to work in a similar field of business may cooperate with the entrepreneur in the new business. This is referred to as transfer of social capital from one context to another. Finally the entrepreneur may engage in forming new relationships with suppliers, customer and clients and even supporting agencies in order to start the operations. The dichotomy between the entrepreneur s social capital and that of the firm tends to overlap as the firm becomes operational. The NV relies on two types of relationships that provide it with crucial resources needed to start the operations and for the establishment of the firm. The first type consists of weak ties, which are often low cost information sources, require less time to develop and provide access to valuable complementary knowledge (Tsai and Ghoshal, 1998). These loosely tied relationships are dominated by business relationships. The second set of relationships is characterized by ties that are deep and strong, in which the actors tend to interact and communicate frequently (such as ties with family and friends). These relationships are often referred to as closed networks (Coleman, 1988; Granovetter, 1983; Harryson et al., 2008). 118

138 Strong relationships are resource consuming and costly to maintain and develop, as opposed to weak relationships. They are very important, however, especially in the case of individuals enjoying high power status (Han, 2006: 103) who have access to highly embedded (tacit) knowledge vital to the NV. The optimal combination of productive relationships would be several weak relationships and few strong relationships (Han, 2006: 104). Open networks are more important during the early internationalisation stage, where weak relationships offer new opportunities for searching and accessing complementary knowledge and resources. Weak ties play a significant role in speeding up the development time of the NV in the foreign market, while few strong relationships are important to provide access to more complex tacit knowledge (Harryson, Dudkowski and Stern, 2008). For instance, an entrepreneur may be given a reference to a local contact in a particular market through a friend, who is knowledgeable about that market. This is a cheap source of information and does not require much time and effort to maintain. However, if the host country is known to be a very risky business environment, a strong relationship with a local partner or a client may offer the entrepreneur knowledge that is confidential and highly tacit on how to overcome such risks. Therefore, weak relationships offer NVs market knowledge necessary to enter foreign market at relatively low cost, while a few strong relationships are required especially if highly complex and tacit knowledge is essential. Contextual effects The NV is embedded within its local context (home country) and this has an impact on the speed of internationalisation. The institutional context tends to shape "the rules of the game" (North, 1990:1) and define rules and regulations that foster and impact the entrepreneurial activities and outcomes. In addition, the culture of home country has an impact on the extent 119

139 to which the entrepreneur values relationships when doing business (Kiss and Danis, 2008). For instance, in countries like China relationships are crucial for doing business in contrast to western societies where the business tends to be more important than relationships. That does not necessarily mean that relationships are unimportant, but is does mean that people can do business with each other even after just one meeting. Accordingly, the institutional context and culture of the home market has an impact on the speed and outcomes of internationalisation as well as on the extent to which the entrepreneur relies on different types of relationships. Early Internationalisation stage [EIS] Oviatt and McDougal (1994) proposed that new ventures usually internationalise directly after their inception or soon afterwards. Empirical studies have normally used six years as a cut-off point such that a company is said to be a NV if it internationalises within the first six years of its operations (Oviatt and McDougall, 1994; Arenius, 2002). The initial or first market entry could be triggered by the presence of a critical event (E2 in Figure 3.1). This critical event could be based on a rational decision of the entrepreneur to internationalise, having engaged in an active search and exploration of new markets (Barkema and Drogendijk, 2007). Or it could result from a serendipitous event such as an unexpected opportunity arising in a particular market, whereby one of the company s client moves into a new foreign market and the company follows him or whereby the entrepreneur meets someone at a business dinner or at a social club who proposes a business opportunity (Crick and Spence, 2005). Accordingly, SME entrepreneurs will assess their portfolio of relationships and the growth strategy of their firm (Han, 2006). They will then engage in developing new relationships (within the local context or in foreign markets) that are needed to provide the firm with the 120

140 various resources and information needed for its growth or they may redeploy an existing relationship from one market to other similar markets. For instance, if a sales agent in an Asian country proves to be a successful partner, the firm may ask him to be responsible for assisting entry into a nearby country that has similar market conditions. In this case, the entrepreneur would focus on the further development of existing relationships which retain the potential to increase the firms growth and survival in their respective positions (markets). Over time the initial network social capital will change and evolve, with the weak ties becoming stronger and well established because of repeated interaction (Adler and Kwon, 2002). Furthermore, the more the firm engages in international operations and expands into new markets the more it develops business ties. This would create for the NV an evolving network social capital. It may be composed of multiplex relationships composed of both business and social ties (Larson and Starr, 1993). The subsequent Internationalisation Stage [SIS] The subsequent internationalisation stage involves further internationalisation of the firm and is usually characterised by extensive international involvement and perhaps the use of more advanced modes of entry, such as joint ventures. This is usually an intentional decision taken by SME decision makers to aggressively internationalise. This could be a proactive decision or as a result of an external stimulus (i.e. reactive decision). With the passage of time, if an opportunity or threat (critical event E3 in Figure 3.1) emerges from the extra-institutional context of host countries, the NV is forced to consider expanding into more distant markets as well as to increase its involvement in existing markets. Thus, it is likely to focus on securing its market position in existing markets (exploitative 121

141 internationalisation) and expanding its knowledge base about these markets (exploratory internationalisation) 17. As the firm becomes more established in international markets, it becomes a global player focusing mainly on exploiting its well-developed knowledge base to increase its market share in existing markets, with minimal attention given to further exploration of new markets. It is a fact of life that relationships decay, grow and change (Prathanham and Dhanaraj, 2010). The NV s network social capital will change and evolve with the evolution and internationalisation of the NV. This evolution can occur to both structural capital and relational capital. Structural capital evolution Structural capital evolution can occur at both the overall network level and at the individual dyadic level. Structural capital evolution at the overall network level The inter-firm relationship literature has highlighted the fact that firm s social capital may increase or decrease over time (Doz, 1996; Ring and Van de Ven, 1994; Prashantham and Dhanaraj, 2010). In her study Coviello (2006) found that NV network social capital increases over time as the network s size and density increases. Its network increases in size as the NV 17 Exploitative internationalisation refers to the case where the firm expands into nearby and low psychic distance markets, where the firm can exploit its resources and current knowledge base about a country s institutional context this includes rules, procedures and standards - and culture into a similar context. This involves expanding and enhancing the company s current knowledge base in a similar context. Moreover, internationalising SMEs could exploit their current technical and market knowledge (related to their products and services) to similar contexts and this is another form of exploitation (Barkema and Drogendijk, 2007). By contrast, the firm may internationalise into psychically distant markets. These markets have new ways of doing business, different cultures and institutional contexts. Accordingly, the firm needs to actively engage in searching for new knowledge and opportunities in these markets through "the process of planned variation and experimentation and is more uncertain in terms of its outcomes" (Barkema and Drogendijk, 2007:5; Baum et al., 2000). Thus, exploratory internationalisation involves entering into far and distant markets, this involves active search for new sources of information and opportunities through the process of planned testing and disparity. 122

142 engages in forming new relationships. Firstly, the SME will form relationships with supporting agencies such as its local chamber of commerce and export promotion agencies. Secondly, it will form relationships with clients and suppliers (value-chain networks) through joint and collaborative agreements (Andersson, 2000; Ruzzier et al., 2006). Finally, the NV resorts to forming relationships with distributors in foreign markets. Relationships with distributors are the form of network connection most discussed in literature due to their strategic position as main entry paths into entering new markets. A high degree of trust and commitment needs to be established and nourished with them. Accordingly, the NV becomes more established and its legitimacy increases. The overall network will evolve following either of the following scenarios: the whole network will evolve into a sparse, weak configuration dominated mainly by business ties; or it will evolve into a closed network with strong and deep relationships based on personal and social ties. These alternatives are expressed in the differing views of Hite and Herstley (2001) and Larson and Starr (1993). Hence, it is expected that the NV will end up with different types of networks having varying degrees of strength. Evolution of dyadic relationships It is a fundamental assumption of the present study that tie evolution can take a wide variety of forms. As a firm evolves it will establish and expand its network social capital through the creation of new relationships (Prashantham and Dhanaraj, 2010). This is crucial for the firm in order to gain access to the market, together with technical and other support required for it to succeed in foreign markets. However, not all ties remain the same and the NV s network social capital could decrease or depreciate over time (Adler and Kwon, 2002). This is evident in Burt s (2002) exceptional study which focused on bridging ties that ceased to exist within 123

143 a year s time. In other words, the strength and value of tie may diminish with passage of time" (Soda, Usai and Zaheer, 2004:893; Prashantham and Dhanaraj, 2010). The reasons for the depreciation of social capital can range from unlawful acts, lack of commitment, tie decay, and reduction in tie utility (Burt, 2002; Prashantham and Dhanaraj, 2010). A tie decays (ceases to exist) if it diminishes, becomes more fragile and has not been used for some time (Burt, 2002; Bourdieu, 1986). It can result from the focal actor becoming preoccupied with other actors (Burt, 2002; Putman, 2002; Prashantham and Dhanaraj, 2010) Another possibility is reduction in tie utility even to the point of obsolescence. A certain tie may not produce or generate any further business opportunities (Prashantham and Dhanaraj, 2010). In their study Prashantham and Dhanaraj (2010) found out that with passage of time some ties between the software companies (focal firms) and some foreign clients could not yield any further growth opportunities, which meant that they failed to meet the expectations placed by focal firms on these key actors. If a tie becomes obsolete, this means that its effectiveness is highly dependent on a particular context (Jack, 2005). Obsolescence could be due to changes in the context of the relationship such as when an actor retires or leaves his position. While there may be a strong relationship between the focal firm and this actor, it is no longer beneficial because his circumstances have changed. In some extreme cases, the government of the foreign country may prohibit the production and/or sale of certain goods. If the NV sells this product through a distributor, this development will force the NV to retire this business relationship because the reason for the 124

144 collaboration has ceased to exist (Helfat and Peteraf, 2003). This could happen even though the relationship between the focal firm and these actors is still being frequently enacted and is valued by the focal firm (Ring and Van de Ven, 1994). Finally, the NV may actually decide to retire or retrench some ties for other reasons. A learning race may develop between the firm and others, in which case the latter become a threat rather than a support (Hamel, 1991). One of the focal firm s partners may demonstrate lack of commitment by failing to meet his/her obligations. A third possibility is that a large client may place restrictions on the NV, such as an MNC on which it is "piggybacking" for its foreign sales (Coviello and Munro, 1995). In some cases there might be a decline in customer demand or the emergence of new competitors, which could force the NV to shrink its operations, thereby reducing the number of its business relationships the retrenchment of existing relationships to a few suppliers and/or customers. Accordingly, the NV would focus on strategic relationships and reduce (divest) those relationships that are less important. In addition to tie appreciation and depreciation, or the expansion and contraction of ties, there are other patterns that could take place on dyadic level. The focal firm (NV) may recombine several relationships to create a specific network, redeploy or reactivate existing relationships, and keep relationships in dormant mode until needed. New relationships may need to be established with key customers, suppliers or distributers which could be crucial for the NV s growth and survival. The NV may reactivate past relationships in order to seize an opportunity that arises. For example, an entrepreneur may know someone who is an expert on the Brazilian market this expert used to work with the entrepreneur but the contact was lost because the entrepreneur opened his own business. When an opportunity arises for the 125

145 company to enter Brazil the entrepreneur may reactivate this relationship in order to access knowledge about the market. In the case of restrictions on the production of a certain products or components, the NV may redeploy or reuse this specific relationship to access other niche markets within the same country or enter new markets. For example, the government may prohibit use of a certain component of a final product, in which case the NV may use the business relationship with this supplier to produce another component to meet another target market. Another possibility would be to redeploy the contact or business relationship to enter another foreign market. Finally, in times of crisis, the NV might form inter-firm collaborations with its key suppliers and/or competitors, responding to challenges in the business environment through forming a specialized network that has a portfolio of complementary resources enabling it to overcome emerging threats. Another possibility is that the emergence of new technology might oblige the NV to form a collaborative agreement with some of its partners in order to make use of such new technology and improve its profitability. Some of these options could follow one another. For example, at a time of crisis some of the potentially profitable relationships might be kept on hold, whereby minimal effort and time is required to maintain them. Later on, those relationships could be reactivated to realize opportunities or to help overcome prevailing threats. In other words, letting existing relationships lie dormant might be followed by their reactivation. In similar vein, existing 126

146 relationships might be kept to a minimum (retrenchment) for a period of time, and then intensified again. It is sometimes not possible to connect directly with a particular contact, but this has to be done through creating a bridge relationship between them (Hakasson, 1992). For example, a British company wants to form a partnership with a key Chinese supplier. However, negotiation with this supplier has proven to be very difficult and a consensus on managing this venture is problematic, due to different leadership styles. The British company approaches a key customer of its Chinese supplier and forms a cooperative technical project with that customer. This helps the British company to approach the supplier through an intermediary or bridging relationship. Another bridge could occur if a key employee of the British firm has previously worked with an established client of the Chinese supplier. These are examples of bridging relationships that an SME could create to connect to networks they do not have direct access to. Such networking development patterns differ from one situation to another and the way they evolve is likely to depend to a great extent on several factors such as the history of the relationship, the perceived benefits for each partner (reciprocity), the costs inherent in developing and maintaining such relationships, and external factors within the NV's institutional and economic contexts. These factors are assumed to have a moderating effect on the tie development process. 127

147 Relational social capital development With the passage of time, frequent interaction and meeting expectations of different actors within the network enhances the level of trust and mutual commitment. In addition, positive reinforcement on the part of NV can repair and maintain those ties that are weakened or are problematic. A high level of trust and mutual commitment between partners may lead to exchanges that often go beyond what is specified in formal contracts, and it helps to reduce the risk of opportunist behaviour by the partners (Kilpatrick and Bell, 1998). In addition, trust eases the transfer and acquisition of tacit knowledge that is crucial for the firm s survival. All in all, over time the NV is expected to enhance its internationalisation learning process and its networking capability which in turn should lead to improved sales performance and growth at the present time (Tn in Figure 3.1.) This is reflected in feedback loops in Figure 3.1. Thus, both the NV s internationalisation and network evolution processes evolve together in such a way that they influence one another, giving rise over time to better firm growth and survival through adaptation to changes in the external environment. Nevertheless, the co-evolution of internationalisation and networking does not necessarily lead to improving performance if, for example, the internationalisation proves to be ill-judged or the choice of a problematic overseas relationship (for example, an agent) which might cause the company to de-internationalise. In addition, the NV evolves with its external environment by reaching accommodations with that environment. Its leading entrepreneur may be able to use her or his network relationships to lobby or negotiate the enactment and enforcement of laws and regulations beneficial for the firm. In addition, s/he may collaborate with supporting agencies and obtain beneficial 128

148 forms of support. The dotted line at the top of Figure 4.1 denotes the co-evolutionary frontier between the NV and the external environment. Figure 3.1 also highlights a cyclical loop where each of the above mentioned processes influences and evolves with one another and with the external environment. Not only does the NV adapt to changes in environment through responding to critical events but also the firm in some cases through its position in network can influence the external environment and trigger some changes that have future effects on firm s growth and survival in the long run. The traditional SME co-evolutionary model The traditional SME tends to internationalise through incremental and gradual steps following the traditional process models of internationalisation. SMEs may not pass through all the stages described in the Uppsala model; it may even skip steps (Barkema and Drogendijk, 2007; McGrath, 2001). However, it will most probably start off with supplying to the domestic market, and then expand to nearby low psychic distance markets (especially if these SMEs come from emerging or developing economies). Then, as its level of experiential knowledge increases, it starts internationalising into further-distance markets. Also, it may use multiple entry modes such as licensing, joint ventures or even forming local subsidiaries in foreign markets (Johnson and Mattsson, 1988; 1985). Another possibility would be, once SME s home market becomes saturated, the only option for SMEs is to expand overseas. At this stage the SME would explore new markets in far and distant markets as well as exploit its current knowledge base in nearby markets. These two actions occur concurrently. This is likely to 129

149 depend on the type of product and industry in which the firm operates the level of experience the firm has gained from its domestic operations and the local context in which it is embedded. These propositions are depicted in Figure 3.2. The local institutional context in developed markets tends to foster and encourage entrepreneurs and firms to internationalise through offering needed sources of support (Kiss and Danis, 2008). Also, the presence of mature institutions in these markets encourages entrepreneurial behaviour and firm internationalisation more than is likely to be the case in less developed markets (Dodd and Patra, 2002). However, if the local institutional environment is restrictive this could encourage the SME to internationalise as a more attractive outlet for their business (Boisot and Meyer, 2008). Start-up and domestic expansion stages Accordingly, during the start-up stage the firm follows the same scenario as in the case of NV when forming the business, but it tends to rely mainly on strong local ties and few weak domestic ties. As time passes, the SME s initial network social capital will evolve in a way that helps the firm to expand locally. The firm s network social capital evolves into a closed network (composed of both strong local business and social ties). International expansion stage The firm becomes well established and well known in domestic market and perhaps in nearby markets. A critical event might then arise in the firm s extra-institutional (market) environment that leads the SME to internationalise. This could be a new opportunity in a foreign market, including a serendipitous event. Or it could be a pressure to go abroad such 130

150 as domestic market saturation. It is likely that traditional SMEs would start by exporting to low psychic-distance overseas markets (Welch and Luostarinen, 1993; 1998; Calof and Beamish, 1995; Johanson and Mattsson, 1993). As those SMEs learn more about overseas markets and gain confidence, they would start to enter more distant markets and perhaps use more advanced investment-based modes of entries (such as joint ventures or wholly owned subsidiaries). This suggests that the traditional SME might engage first in exploring similar overseas markets and that once their level of experiential knowledge increases they would exploit this knowledge and use it in other overseas markets (Barkema and Drogendijk, 2007). Thus, the traditional SME entrepreneur would engage in both explorative and exploitative internationalisation. If the firm consciously planned to internationalise or extend its existing foreign involvement, it would then assess pre-existing relationships (if any) in local and foreign markets that could offer access to required resources and support. In addition, it would need to develop new relationships in those markets and with relevant supporting agencies (such as export promotion agencies) in order to gather acquire needed market knowledge. The SME may transfer or redeploy an existing relationship from one context to another. For example, a friend or a relative of the entrepreneur may have helped the entrepreneur in starting his business and later they may agree that this relative would represent the company in another market. Over time the internationalisation level of SME increases, as well as its international experience and level of experiential learning. This may encourage the firm to increase its level of international commitment through using several modes of market entry. 131

151 The SME s network social capital is seen to evolve in the same way as suggested previously for the NV's subsequent internationalisation stage [SIS]. The structural social capital evolves at both the overall network level and that of individual dyadic relationships. Overall networks develop into more sparse and multiplex networks composed of both strong and weak ties (social and business ties). The individual dyads will also follow the above mentioned patterns of tie development: some ties may depreciate, appreciate, enter into dormant mode and then become reactivated or recombined (Jack, 2005). Relational social capital will evolve. Increased frequency of interaction and positive outcomes, lead to an enhanced level of trust and commitment between the focal firm and its partners. Over time, the SME's level of learning increases in terms of network learning and experiential learning, giving rise to an enhanced level overall performance and growth at Tn. The SME will also co-evolve with its environment through responding to changes inherent in both its local environment and in its extra-institutional market environment (Lewin and Volborda, 1999). It will also interact with its institutional environment through initiatives taken by its entrepreneur to change certain laws or circumstances which have a bearing on the firm s future growth and expansion strategies (Child, Tse and Rodrigues, 2013). The possibility of securing support for such initiatives depends on whether the entrepreneur occupies a structural hole or has access to structural holes. In short, Figures 3.1 and 3.2 identify somewhat different co-evolutionary processes, depending on the type of SME new venture or traditional. 132

152 Conclusion As explained in Chapter 2, the present research aims to understand the co-evolution between SME internationalisation and network processes over time, and to compare the coevolutionary patterns of UK and Egyptian SMEs. The UK is an example of a developed economy with a strong institutional context and Egypt is an example of a weak emerging market with a weak and inefficient institutional context. I propose two different models of co-evolution pertaining to traditional new ventures and traditional SMEs. The literature reviewed in Chapter 2 indicated that each of these two types of firm is likely to follow a different path of internationalisation. New ventures networks may start off with a set of interconnected social relationships. Over time their network is likely to be composed of both weak and strong ties and it will evolve from being a identity based, path dependant network to a deterministic/path dependent intentionally managed one (Hite and Hesterly, 2001; Coviello, 2005; 2006). By contrast, traditional SMEs may start off with a one-dimensional set of relationships comprising either social or business relationships. This evolves into a multifaceted combination of both social and business relationships (Larson and Starr, 1993). These types of ventures or SMEs are likely to be present in the two countries Egypt and the UK in which they are to be sampled. However, since each country differs in terms of culture and level of institutional development, the networking behaviour of their SMEs is expected to be different in ways also indicated in Chapter 2. The generic framework I am proposing allows for variation in the networking behaviour of the two types of SMEs from both countries. That is why I have not proposed a particular network social capital development pattern but rather used generic terms such as initial network social capital and 133

153 evolving network social capital (composed of both structural and relational social dimensions), which allows for an empirical investigation of variations in patterns of network development. Moreover, it is important to note that the SME lifecycle stages differ in the two models. The NV model in Figure 3.1 proposes three stages namely (1) the pre-internationalisation stage [PIS], (2) the early-internationalisation stage [EIS], and (3) the subsequent internationalisation stage [SIS]. By contrast, the traditional SME s model proposes three somewhat different stages to allow for a different internationalisation pattern. These are (1) the start-up stage, (2) the domestic expansion stage, and (3) the international expansion stage. To recap, both co-evolutionary models described in this chapter offer a general overview of the likely co-evolutionary patterns of NVs and SMEs from both countries. The models are a priori in nature having been derived from findings and discussions in the existing literature. The intention is later to formulate a revised model based on my empirical findings which will offer a more precise understanding of SME co-evolution. 134

154 CHAPTER 4 METHODOLOGY Introduction This chapter discusses the methodology used in this research. It starts by discussing the theoretical orientation of the research, followed by a description of the qualitative research strategy and methods used in data collection. Finally, the process through which data was prepared and analysed is presented 18. Theoretical Orientation This study combines two theoretical orientations. The first relates to the fact that this research deals with a phenomenon that draws heavily on the real experiences (in terms of internationalisation and networking behaviours) of social actors (SME decision makers). Some scholars in the international entrepreneurship field argue that understanding the real experiences of entrepreneurs requires the researcher to adopt an interpretivisit or phenomenological stance (Lamb, Sandberg and Liesch, 2011; Coviello, 2005). A phenomenological stance suggests that understanding the real experiences of social actors, should be based on the interpretation and perception. Since internationalisation and networking behaviours are idiosyncratic to entrepreneurs, they are better understood through the eyes of those actors who created them (Lamb et al., 2011). On the other hand, the mainstream SMEs internationalisation literature adopts an economic perspective on internationalisation and suggests that the degree of SME s internationalisation 18 Several parts of this chapter were submitted as research methods assignment during the author s first year of PhD studies (Narooz, 2009). 135

155 can be understood through focusing on a few pre-defined variables. In a similar fashion, the networking behaviour of SMEs has been viewed by some scholars from a structural point of view, reducing this to a set of structural/quantitative variables (c.f. Coviello, 2006). One of my research objectives is to investigate how the internationalisation and networking behaviours of SMEs co-evolve with one another over time. The only way to investigate the correlation between these two processes is to rely on codified variables (representing structural network and internationalisation performance measures) at different points in time and to test if there is a correlation between the two processes using statistical techniques. This meant that I had to adopt methods and an orientation that approximates to the natural sciences, in other words, a positivist stance. This inherent tension regarding which philosophical position I should take evolved with the different stages of research. I started with an interpretivist position, viewing the world as purely socially constructed, since I am dealing with the real life experiences of entrepreneurs. However, through the process of data collection I began to realise that I was being drawn towards a positivist view, because the evolutionary patterns of SMEs could only be traced through narrowing down my focus to measurable outcomes of internationalisation and network structures. However, as I became more immersed in analysing the data I could see that if I only focused on these few measurable variables I would lose the important and rich qualitative insights into the reasons why internationalisation and networking developed together; insights that each entrepreneur expressed through his or her own interpretation were highlighted. 136

156 Eventually, I realised that I am neither adopting a purely positivist nor purely interpretivisit stance, but a position that lies somewhere between the two. The theoretical stance that I felt close to is the critical realist position suggested by Margret Archer (1995; 2003). According to Blundel (2007), this helps researchers to cross the bridge from the natural sciences, with its focus on objectivity, to the social sciences with their recognition of the role of perception and interpretation. The following section describes the epistemological and ontological orientations consequently adopted in this study. Epistemological and ontological orientations Methodology has to do with how the researcher caries out the research and it is clearly influenced by the way the researcher views reality, either as an independent entity or socially constructed by actors (Ryan, Scapens and Theobald, 2002; Potter, 2000; Johnson and Duberley, 2000; Descartes, 1639; 1641). The way the researcher views reality represents his or her ontological orientation. Also, the research process is influenced by how researchers verify that their knowledge of reality is true and acceptable, and that the sources through which knowledge is obtained are valid. This is referred to as the epistemological orientation (Ryan et al., 2002; Potter, 2000; Bryman and Bell, 2003). Accordingly, it is important to identify the appropriate epistemological and ontological orientations to guide the research methodology and methods chosen to investigate a particular phenomenon. I hold the view that the reality is not purely an independent entity but it is also perceived and constructed by the social actor s experiences and descriptions. For example, the internationalisation of SMEs embedded in a particular institutional context (Egypt) is influenced by same set of challenges and dysfunctionalities, which are real independent 137

157 circumstances. However, one entrepreneur could perceive these dysfunctionalities as an opportunity that he can change and influence to his advantage. Another entrepreneur might perceive these challenges as a major obstacle for his firm s growth and survival. Here each one of them holds a different view about the challenges inherent in the Egyptian context and a different way to respond to it. It is an established fact that the Egyptian context is full of institutional voids, but whether we see these as a challenge or an opportunity is socially constructed by our experiences and perceptions of this real independent condition. Moreover, I tend to hold the epistemological view that not only is entrepreneurship (in terms of internationalisation) an individual act, shaped by the decision-makers choice of which path to follow, but that there are also certain observable patterns of such actions that could be investigated to provide an overview of the generic behaviour of entrepreneurs. Moreover, entrepreneurs tend to respond to changes in external environment that are imposed on them and to some extent this determines the way they will respond to these external stimuli. Accordingly, my perspective brings together two epistemological orientations: positivist and interpretivist. These alternatives can be understood by reference to Burrell and Morgan s (1979) representation of the assumptions underlying nature of social sciences, see Figure

158 Realism In essence, social and organizational reality exists independently of human consciousness and cognitions. Ontology Epistemology Nominalism Reality is simply a product of our minds, a projection of our consciousness and cognition with no independent status Positivism it is possible to observe the empirical world in a neutral manner through the accumulation of objective sense-data Determinism Human nature Anti-positivism It is not possible to observe the empirical world in a neutral manner through the accumulation of objective sense-data Voluntarism Sees human behaviour as determined by the situation, as necessary responses to external stimuli Nomothetic Located in the unity of the sciences and applies protocols and procedures derived from the natural sciences Methodology Human action arises out of the culturally derived meanings they have deployed during sense-making Ideographic Attempts to uncover the internal logics that underpin human action by deploying methods that access cultures *Source: Johnson and Duberley (2000: 78) Figure 4.1: Burrell and Morgan s (1979) assumptions underlying nature of social sciences 139

159 So rather than being at one of the extremes in the four dimensions highlighted by Burrell and Morgan (1979), my position lies somewhere in the middle. As mentioned earlier, the closest theoretical orientation that allows for this combined view is critical realism (Bhaskar, 1975; Archer, 1995; 2003). Bhaskar argues that "what we can know sets the ground of how we know something exists" (1978:36). At the heart the critical realist view is that social scientists cannot be absolute subjectivists (relativist) or absolute realists, but rather they are transcendentalists. Critical realists argue that there is some middle ground between the positivist approach which follows a reductionist approach and tends to pay little or no attention to the depth and richness of social phenomenon and the intrepretivisit approach which argues that the social phenomenon can only be understood through the eyes and perceptions of social actors. Furthermore, critical realists suggest that it is important to allow room for social constructivism and at the same time to understand that reality is independent from the researchers/observers. More recent critical realists such as Archer (1995; 2003) suggest adopting a morphogenesis 19 approach to understanding the nature of social reality. At the heart of Archer s (1995) argument is that structure is a social reality which exists at a given point in time. While it may have been created by agents previous actions, in the present time it has properties and exercises an influence that is independent of the agent. However, what agents decide to do have implications for future structures the agents may to some degree be able to shape future structures (Archer, 1995; 2003; Mole and Mole, 2010). 19 Morphogenesis according to the Oxford Dictionary is the branch of biology that studies the forms and structures of living organisms. 140

160 She holds the view that structure exists as an objective reality that materialises as a result of actions of past agents, it is a reality that is recursive and that it has "causal power that are independent of agents" (Mole and Mole, 2010:235). These emergent structures could constrain or empower agents. Agents are spontaneous and fallible they could make right and wrong choices and they learn from past experiences. In other words, from an ontological point of view Archer argues that social reality exists independently from the agent and can be observed. It is also a product of the agent s past action and over time it materialises and has its own properties that are independent from the agents (Archer, 1995; 2003; Mole and Mole, 2010). Moreover, we can observe this social reality as a structure and also we can understand the agents and their actions and the interaction between the structure and the agent. Archer builds on Burrell and Morgan s (1979) human nature assumption that agents are influenced by stimuli from the external environment and that they respond to these changes but also they are capable of making decisions that might change future structures and sometimes even shape it (Mole and Mole, 2010). She suggests that to better understand the interaction between the structure and agent a dualism analytical lens should be used 20. She argues that agents and structures change over time. Sometimes the structure determines the agents actions and in other times it is the agent who shapes or respond to these changes in structure, the interplay between them moves in a cycle and neither agents or the structures has precedence over the other (Mole and Mole, 2010). 20 This is different from Gidden s (1979; 1984) duality concept. He argues that structure and agents are inseparable and cannot be understood independently from the other. 141

161 In methodological terms, Archer (1995; 2003) argues that in order to understand the interaction between agents and structures, qualitative and quantitative methods could be used to capture the objective observable reality of structure and at the same time highlight the agent s choices and roles in responding and perhaps shaping the future structure. Archer s critical realist view is similar to the theoretical orientation guiding this research. The first assumption of this research is that one way to understand an SME s pattern of internationalisation through observing their internationalisation performance (in terms of number of markets and regions and percentage of foreign sales) at different points in time. These internationalisation performance measures are structured in nature, structured in patterns, and can be observed to exist at a given point in time. Likewise an SME s configurations of network links as well as the network itself are structured in nature. In other words, the SME s structural network configuration and attributes can be observed to exist at given points in time and structured in terms of network development patterns. The second assumption is that the wider institutional environment is also structural in nature. In other words, it exhibits a structure of rules, norms, organisation and officials that are independent of the SME. The institutional environment is an example of an independent structure that enables and constrains the agent s actions (in this case the SME decision maker). At the same time there are agents (SME decision-makers) who relate to these structures and for whom the structures are meaningful. Moreover, the agents have the ability within certain limits of their knowledge and power to make changes in their future internationalisation and 142

162 their networking structures. Thus the changes in those structures over the course of time are assumed to an important extent to result from the agents actions. For instance, SME decision-makers can decide to enter new foreign markets or withdraw from a foreign market; similarly they can change their network links. Some of the characteristics of the foreign markets and the network others may independently change over time as well. These actions taken by the SME decision-makers-focal actors- can be assumed to constitute the main source of any co-evolution between internationalisation and networking. Since the focus of this study is concerned with the internationalisation and networking processes of SMEs that are managed by agents (decision makers). This critical realist theoretical orientation has three important implications for the methodology used in this study. Firstly, the structured features (internationalisation and network links) can be regarded as potentially observable objective facts. They are therefore amenable to a positivistic research methodology, although in practice it may not be feasible to observe them directly but rather the source of information on them that can be practically accessed, as can actors whose description of the structured features could be coloured by interpretation and rationalisation. The fact that these features are structured means that they are open to classification, codification and quantitative measurement. Secondly, the actors rationales regarding their internationalisation and networking decisions are subjective and interpretative in nature. They are necessarily informed by their understanding of their situation, both in aspects specific to a particular decision and the general situation. Therefore, the appropriate research methodology here is qualitative, which 143

163 allows for interpretation and understanding of the subjective decisions and rationales of SME s decision makers. Moreover, using a qualitative methodology would provide the information on internationalisation and networking behaviours of SMEs that could be codified and quantified. To do this, I adopted a specific methodological approach, whereby I depended mainly on multiple case studies to collect data about the different SMEs experiences. I collected data from forty SMEs in UK and Egypt, which is arguably a sufficient number of cases to enable me to identify key patterns of development in each country. When analysing the data I followed the "bi-focal" or mixed method approach suggested by Coviello (2005). Her approach enabled me to focus on the key internationalisation and network variables and trace their change over time from each sample and compare the behaviour of the UK and Egyptian SMEs. I was also able to test for possible correlation(s) between internationalisation and networking as will be reported in Chapter 5. In addition, I wanted to understand how these two processes evolved and co-evolved over time and what were the rationales that entrepreneurs articulated for the actions they take. This was done through closely examining the individual experiences and interpretations of actors 21 to understand why a particular entrepreneur chose to follow a certain internationalisation path or build a relationship with a particular network contact. These processes could only be extracted from the narratives provided by entrepreneurs as will be reported in Chapters 6 and Although some studies define entrepreneurs as risk takers and are proactively seeking and creating opportunities. In this study the term entrepreneur is used interchangeably with SME decision maker. Both refer to the individuals in charge of SMEs. 144

164 The next section explains the different research strategies and methods used to collect and analyse the data. Research strategy A quantitative research strategy emphasises the use of measurement and quantification of data. Hence, it resembles the use of scientific methods used commonly in the natural sciences. It falls under the umbrella of a positivist epistemological position; and has an ontological orientation that views social world as "an external, static and objective reality" (Bryman and Bell, 2003:25; Gill and Johnson, 2010; Bryman and Bell, 1988). The main advantages of quantitative research are that it enables the researcher to quantify qualitative data which makes is easy to co-relate to other quantitative data, and hence to infer causality; and to generalize the findings of a representative sample to the wider population and replicate the findings in other studies (Bryman and Bell, 2003). Quantitative research is a strategy that involves the use of two main approaches. One is the deductive approach that deals with formulating hypothesis about the cause and effect that would be empirically tested and out of which "explanations of laws and theories are to be assessed" (Bryman and Bell, 2003:14). The other is the inductive approach which deals with gathering facts and data that form the basis for formulating theories and laws. The argument for induction is that only through observing phenomena in real life that explanations and predications can be reached in a valid and objective way. 145

165 Adherents to a quantitative approach are more likely to use research methods that emphasize objectivity, repeatability (transferability in the case of qualitative research) and generalizability (in the case of quantitative research). Although this position tries to simplify complex phenomena into more concrete and measurable laws based on cause and effect relationships, it suffers from the difficulty of controlling for all the factors in the environment of social world as opposed to the natural sciences (Rehan ul-haq, 2005; Bryman and Bell, 2003). Moreover, a quantitative research strategy suffers from key shortcomings which are summarised in Table 4.1. By contrast, qualitative research tends to focus on processes and non-quantifiable data, through the use induction and theory building rather than theory testing. It focuses on the "ways individuals make sense of their social world" from an interpretive epistemological position. Moreover, it belongs to the ontological orientation that views social world as processual and socially constructed by the actor (Bryman and Bell, 2003:25; Bryman and Bell, 1988). Table 4.1: Shortcomings of quantitative research Shortcoming Ignores the fact that any social phenomena could be socially constructed Low ecological validity Fails to identify the underlying mechanisms behind causal relationships Explanation It tends to ignore people's interpretation of their social world. The researcher is detached from the social phenomena and his role is only to observe it from a distance (Johnson and Duberley, 2000) Ecological validity tend to be lower than in qualitative research in the sense that the instruments and measures used by researcher may be his/her own artefacts that does not represent the actual phenomena under investigation (Bryman and Bell, 2003). Quantitative research can help ''infer and identify that a certain variable causes another rather than how this causality is produced by the people to whom it applies" (Bryman and Bell, 2003: 86). 146

166 The main advantage of qualitative research is that it allows the researcher to view the social world though the social actor s own eyes. Also, qualitative research tends to be highly context and process specific, and it places a high significance on "how events and patterns unfold over time" (Bryman and Bell, 2003:296). Finally, this strategy allows for a high degree of flexibility and limited structure which more readily uncovers social actors perspectives without limiting the research to a predefined rigid structure that does not allow for taking into account people s perspectives that researcher has not anticipated a prioi. However, this strategy also is not without shortcomings, which are presented in Table 4.2. Table 4.2: Shortcomings of qualitative research Shortcoming Suffers from subjectivity Difficult to replicate and suffers from lack of transparency Focuses on theoretical generalisations and cannot be generated to the wider population Explanation It could be overly subjective because the findings are based on researcher s own interpretations of social phenomena and the rapport/ relationships the researcher creates with the people being studied (Bryman and Bell, 2003; Lamb et al.,2011) This is attributed to the fact that qualitative research is highly unstructured and that data collection and analysis is a product of researcher's own disposition, which makes it difficult to fully understand how data has been gathered and how conclusions have been drawn (Bryman and Bell, 2003) 22. Focuses on analytical inference of the findings gathered from the qualitative research for theory (Saunders, Lewis & Thornhill, 2006; Ryan et al., 2002). Other important considerations that need to be taken into account while choosing which research strategy to use are validity, reliability and replicability considerations. Validity, reliability and replicability There are three main criteria that are commonly used to evaluate different research methods and designs. Firstly, the reliability of research is concerned with the extent to which the questions or measures used will yield the same results whenever repeated (Kirk & Miller, 22 Although some authors like Yin (2008) argue that if the researcher states in a clear and detailed manner how he/she actually collected his data and reached to the conclusions this increases the reliability and transparency of qualitative research. 147

167 1986). Secondly, replicability criterion is concerned with the extent to which findings of other researchers can be replicated- this may be the case if the researcher feels that these results do not resemble other important empirical evidence related to the same topic in question. In order to, increase the replicability of the research, the procedures used to conduct it should be clearly explained and stated (Bryman and Bell, 2003). The final criterion is the validity of research, which refers to "whether a specific measure of a concept really measures that concept" (Bryman and Bell, 2003:77). There are different types of validity which are set out in Table 4.3. Table 4.3: types of validity Validity Type 1.Measurement/Construct Validity 2. Internal validity 3. External validity 4. Ecological validity Definitions The extent to which the measurements used are measuring what is supposed to be measured The extent to which alternative explanations of causal relationships are explored and taken into account. The extent to which the results of a study could be generalized to the whole population beyond the research context. The extent to which the findings are applicable to the people every day natural settings. In other words these findings are not artifacts made by the researcher himself rather than true reality of what happens in everyday life of people. *Source: Bryman and Bell, 2003; Kirk and Miller,

168 These criteria are mainly used to evaluate quantitative research. Some writers argue that they cannot be applied in qualitative research (Bryman and Bell, 2003). However, writers have used alternative criteria for evaluating qualitative research as described in Table

169 Table 4.4: Alternative criteria for evaluating qualitative research Evaluation criterion Definition How it was applied in this study Communicative validity Pragmatic validity Transgressive validity Procedural reliability Refers to the importance of achieving consistency between the researcher s understanding and interpretations of data and the actual data being analysed (Kvale,1989; Sandberg, 2005; Lamb et al., 2011) Refers to the tendency to avoid face-value and straight forward statements and assertions (as is the tendency in positivist research) and to look for the underlying meanings (Sandberg, 2005; Lamb et al., 2005; Kvale, 1989) Refers to the ability of the researcher to identify key differences and variation in phenomena under investigation rather than looking for abstract general consensus (Sandberg, 2005; Lamb et al., 2011). This involves the use of the appropriate research procedures, design and methods to obtain reliable findings and data (Ryan et al., 2002; Yin, 2009). -This was achieved partially through focusing on melting the ice and establishing rapport between researcher and the interviewees before the interview through exchanging s and follow-up calls to ensure that the respondents had a clear understanding about the research. Moreover, during the interview the researcher allowed some time for establishing rapport through asking few general questions about the interviewee s personal aspiration and motivation to start this line of business. This was followed by use of several open ended questions that allowed the interviewee to express and articulate his or her actual experience (Sandberg, 2005; Lamb et al., 2011). This was achieved through probing the interviewees with further questions and asking them to elaborate more. I started by asking generic questions about for example key important relationships that helped the company to start business, internationalise and so forth. With each one of these ties I prompted the interviewee to tell me more about the history of the relationship, how they knew each other, what sort of benefits they offered the SME and how the relationship evolved. This study has two aims mainly to identify general patterns of evolution and coevolutionary among UK and Egyptian SMEs (Chapter 5) and also to identify key variations and idiosyncrasies in individual SMEs behaviour (Chapter 6). One way of ensuring this is that all the field notes and case study material are well documented through creation of what Yin (2009) referred to as "case study database and protocol" which involves details about interview schedule, case study reports, how data was analysed and reported and which could be made available for review upon request by other researchers (Ryan et al., 2002). A copy of the interview schedule, case summaries, detailed codification of data and an example of details of samples and analysis using spreadsheets are described in appendices

170 Transferability Refers to the possibility of replicating and applying the findings in another context (Lincoln and Guba, 1985; Ryan et al., 2002) The aim here according to Yin (2009:44) is to find "replication logic", which in the context of this study means that if the analysis of the data showed similar patterns of network and internationalisation for similar types of firms this would provide strong evidence for theoretical generalisations to the SMEs internationalisation and network theories. Contextual validity Triangulation of data methods; respondents (Feeding back findings to them); theory triangulation. This involves triangulation of data though the use of multiple informants; multiple data collection techniques, multiple sources of evidence and even alternative theories (Johanson and Vahlne, 2003; Ryan et al., 2002). This research used two types of data triangulation-mainly multiple sources of evidence and multiple informants. Although the main data gathered from personal interviews with the founders or managers of SMEs, the researcher gathered a significant amount of secondary data such as brochures, company s websites and newsletters, periodic reports, project databases, annual reports and ongoing list of action plans. These were used before and during the interview to probe and verify some aspects and also were used alongside the interview transcripts to verify and confirm the key events and timings of projects and markets discussed during the interview. 151

171 So far, I have discussed the different research strategies that a researcher can choose from. However, it is equally important to construct a research design that enables the researcher to meet the research objectives and questions. Research designs are the procedures that the researcher follows to collect and analyse research data (Creswell and Plano Clark, 2007). There are different classifications of research design, for example Bryman and Bell (2003) identified four main research designs; namely cross-sectional, longitudinal, experimental, case-study and social surveys. They treat research design to fit either qualitative or quantitative strategies. However, others such as Creswell and Plano Clark (2007) offered a different classification for those researchers using mixed methods. These are triangulation, embedded, explanatory and exploratory designs 23. It is important to select the appropriate research design for addressing the four research questions which are: (1) What are the different internationalisation patterns/paths of Egyptian and UK SMEs over time? (2) What are the network development patterns of Egyptian and UK SMEs over time? (3) Is there co-evolution between the internationalisation and network development patterns of Egyptian and UK SMEs over time? If so, what are the underlying reasons for the co-evolution between the internationalisation and network development patterns based on the understanding and interpretations provided by the UK and Egyptian SME decision makers? (4) What are the key antecedents/contingencies of SMEs internationalisation and networking behaviours 23 Triangulation refers to use of combination of qualitative and quantitative empirical evidence to understand a particular phenomenon. Using both is argued by Creswell and Plano Clark to overcome the weaknesses inherent in quantitative and qualitative research methods. Embedded design involves using different research methods to answer different objectives (Creswell and Plano Clark, 2007). Moreover, explanatory research design or sequential strategy involves firstly the quantitative methods as main strategy and then use in-depth interviews (qualitative methods) to further investigate key issues and vice versa (Creswell, 2003; Creswell and Plano Clark, 2007). Finally the exploratory or two-step design which means using one type of design (qualitative or quantitative) as the basis for the other. For example, the researcher might conduct several case studies to understand a particular phenomenon and then use the insights generated from the qualitative data to develop a measure to be tested on a larger number of cases (Creswell, 2003; Creswell and Plano Clark, 2007). 152

172 To address these research questions, a research strategy and design is required to give the researcher the flexibility to (1) identify key patterns of internationalisation and networking behaviours of SMEs over time; (2) test if there exist patterns of association- between the two behaviours over time; (3) understand the underlying mechanisms and the processes through which SMEs internationalise and use their networks over time. The first two aspects will involve a high level of abstraction and quantification of data (in terms of focusing on network structural variables and on internationalisation performance measures) to provide an overview of networking and internationalisation patterns of the Egyptian and UK SMEs over time and test for the patterns of associations between the two behaviours over time. The third aspect involves reliance on rich qualitative data that provides explanations for these development patterns and more importantly explaining the process by which SMEs internationalise and network over time. Qualitative research is considered to offer the researcher the flexibly needed to gain in-depth and rich data and at the same time allows for abstraction and quantification of some of the variables needed to identify key patterns of development. In addition, there is a need to identify key patterns of development and to test for possible correlations between the two processes, a sufficient number of SMEs is essential, the use of multiple case studies from UK and Egypt is considered the most appropriate research design. In the next section I shall briefly discuss case study design and then move onto the research methods used in collecting and analysing data. Case study design There is a debate on whether to consider case-studies as a research design or as a method of data collection. According to (Yin, 2009:18) it can be considered as an "all -encompassing 153

173 method; covering the logic of design", data collection, analysis techniques and approaches. Traditionally, the case study has been defined as "an empirical inquiry that investigates a contemporary phenomenon within its real-life context" usually using multiple sources of evidence (Saunders et al. 2006, Yin, 1989:13). However, Yin (2009:18) has extended this traditional definition as shown in Figure 4.2. Case-study is an empirical inquiry that Investigates a contemporary phenomenon in depth and within its real life context basically The boundaries between the phenomena & context are not clearly evident. Copes with technically distinctive situation in which there are a lot of variables other than the data point. Therefore, relies on multiple sources of evidence, with data needing to converge in triangulation manner and benefit from prior development of theoretical propositions that guide data collection and analysis. *Adapted from Yin (2009:18) Figure 4.2: Technical definition of case-study as an all-encompassing method Case study research is best used when the research aims to answer "how" and "why" questions; when there is no need to have control over behaviours or events as in case of experiments; and when it focuses on extant or current phenomena (Yin, 2009). Also, Welch, Piekkari, Plakoyiannaki and Paavilainen-Mantymaki (2011) argue that case studies are not 154

174 only used for inductive theory building but they can be used to serve different purposes such as natural experiment, interpretative sense-making and contextualisation of theory. An example is the use of a case study as a natural experiment to refute or confirm an existing theory as Lipset, Trow and Coleman s classic study (1956) of Union Democracy did in respect of Michels so-called Iron Law of Democracy. In addition, case studies could be used to identify complex causal explanations between variables through isolating them from the broader context of the case and using other case studies to verify the causal relationships identified from the initial cases. Hence, this similar to experimentation design and emphasises a greater explanatory power to case studies. On the other hand, when using case studies as an interpretive sense-making approach, the researcher focuses on building theory through interpreting the experiences of actors to gain a deeper understanding of the phenomenon under investigation. Researchers often choose either to focus on theory-testing theory or theory-building. Welch et al. (2011) added a new application for case studies -what they refer to as contextualised explanation. Contextualised explanation involves understanding "how" and "why" a certain event occurred, through adopting a historical view to understand how a certain event or outcome took place through going backwards in time to understand the causes of such event. Furthermore, an investigation can be confined to a single case-study or extend to multiplecase studies. The latter is often referred to as a comparative research design (Yin, 2009; Bryman and Bell, 2003). Using a single case study may place the researcher at risk if the chosen case turns out to be different than what it was expected to be. Therefore, it is crucial to carefully investigate and make sure that the chosen cases match the objective of research to 155

175 minimize this risk (Yin, 2009). There are five rationales for using a single case-study approach as shown in Table 4.5. Table 4.5: Rationales for Using Case Studies Rationale Description -Critical case -Extreme /unique case -Representative/typical -Revelatory case -Longitudinal case The researcher has a set of clearly defined hypothesis prepositions and choosing a critical case that meets all the conditions for testing the theory under investigation which would either refute or confirm theory. That has a unique characteristics that is worth documenting and analysing Where the researcher may choose a case, be it a project, individual or an organization, that is typical of the several other projects, individuals or organizations. When the investigator has an opportunity to observe and analyse a phenomenon previously inaccessible. Studying the same case at two or more different points in time; to trace the changes in factors over time. * Source: Yin, 1989:44; 2009:48 Multiple case studies involve the use of the same methodology as in the single case-study research on two or more cases. This is done through employing "replication design" -similar to using several experiments ensure that vigorous findings are achieved when replicating the findings of first experiment to a second and third one under the same conditions used in the first experiment - with the aim of having more vigorous and strong research findings (Yin, 2009:54). Yin (2009:54) asserts that replication can be either "literal or theoretical replication". The former is achieved through careful choice of each case so that it is expected to yield similar results. The latter is achieved through choosing cases that are expected to have contrasting results but based on predicted justifiable reasons. 156

176 Multiple cases are often used to construct a comparison between cases in order to better understand particular phenomena and to identify whether or not a theory will prevail under different circumstances and conditions (in other words this design enhances the generation of theories). This also helps the researcher to identify which concepts are relevant to existing theories (Yin, 1984; Bryman and Bell, 2003; Eisenhardt, 1989). However, it tends to distract the researcher from the specific context and tends to focus more on how to make comparisons between cases. As discussed in Chapter 2, this research contrasts SMEs from two different contexts and institutional environments, namely UK and Egypt. The latter is an example of a collectivist culture and a weak institutional context and the former is an example of an advanced and well developed institutional context and an individualist culture. These national differences in terms of culture and level of institutional development are expected to shape and influence the SMEs internationalisation and networking behaviours. Hence, use of a comparative multiple case study design would enable the researcher to capture and understand differences in the UK and Egyptian SMEs behaviours. This involves the use of multiple case studies from each context and comparing and contrasting their internationalisation and networking behaviours and the impact of national institutional context of the SMEs behaviour. However, case studies have some key advantages and limitations which are presented in Table 4.6. Table 4.6: Advantages and limitations of case studies Advantages Allow for in-depth investigation of phenomena under investigation Allow the use of both quantitative and qualitative methods. Explanation Allow for addressing a large amount of variables as opposed to surveys where the researcher has only a limited set of variables to focus on. Also, they allow for the use of several sources of evidence such as interviews, documents, and surveys, which enhance the triangulation of data (Yin, 1989). Case studies could involve both of quantitative and qualitative methods. Moreover, it could be used to serve different purposes as a part of mixed methods research design (Yin, 2009; Bryman and Bell, 2003; Creswell, 2003). 157

177 For example, a study may start with few case studies in order to identify key variables and relationships to be tested later with a large-scale survey. Or a study may be primarily qualitative study in one or two organizations and a survey may be conducted among different employees in different departments. In addition case studies can employ several levels of analysis (Eisenhardt, 1989; Yin, 1984). Can be used to address different research objectives Limitations Limited external validity Potential researcher bias for Low internal validity Threats to the reliability of the case(s) It could be used in descriptive research aiming only to provide a detailed description of one or more focal units. Also, it could be used to test a theory using a deductive approach and to building a theory using an inductive approach (Eisenhardt, 1989; Yin, 2009; Bryman and Bell, 2003). This means that the findings of a single case or multiple case studies cannot be generalized to a wider population. Advocates of case study research argue that they aim to achieve "theoretical or analytical generalization"; where the empirical results are generalized to wider theory (Bryman and Bell, 2003:55; Yin, 2009:15). Due to unsystematic and unstructured procedures inherent in case-study research the researcher is more likely to incorporate his own judgements or views into the findings and conclusions (Yin, 1989; 2009; Bryman and Bell, 2003). Where controlling other explanatory reasons for why a certain variable causes another is difficult to achieve (Bryman and Bell, 2003; Kirk and Miller, 1986). However, Yin (1989; 2009) argues that case study can offer insightful evidence that complements results obtained from experiments; this is due to the fact that how and why X caused change on Y is better addressed by case studies rather than experiments, especially if the latter are conducted outside a natural setting. The reliability of a case study is generally low due to its qualitative nature. In other words the ability of the researcher to reach to the same results if he/she carried out the same research with the same case to reach the same conclusions is difficult (Yin, 2009:45). However, there are some techniques that might enhance the reliability of case studies suggested by Yin (2009) which are the use of case study protocol and case study database -where both include details about every stage of research that is documented in great details. As mentioned before, the case study enables the researcher to rely on multiple sources of evidence using different research methods. Research methods refer to techniques used for collecting data such as interviews, questionnaires (Bryman & Bell, 2003).There are several methods for collecting and analysing data ranging from documents (secondary data); different types of interviews; surveys; observations and focus group interviews. So far this chapter has discussed the most common research approaches/strategies and methods. However, there are some specific research approaches/strategies that are unique to the study of relational data and networks. This research approach is referred to as Social 158

178 Network Analysis (SNA), which is an approach used to study relational data and structure of networks using either quantitative or qualitative methods or both. SNA is an approach that has been used by sociologists to study social relationships between different actors as opposed to studying attributes of the actors themselves (Burt, 1978; Hanneman, Robert and Riddle, 2005). The unit of analysis is usually the social network which is a network of actors and relationships between them (Wasserman and Faust, 1994; Edwards, 2010). These actors could be individuals, organizations or groups of people which constitute a set of nodes. The relationships between them are represented by lines between those nodes. There are two traditions in the collection of data about social relationships. The first involves the use of quantitative methods to collect data for SNA such as surveys, measures and maps of relationships in order to grasp the structural dimensions of networks using complex statistical techniques (Carrington, Scott and Wasserman, 2005; Edwards, 2010). The second, less dominant approach, involves using of qualitative methods to obtain information about social relations. These methods use ethnographies, in-depth interviews, observations and mapping techniques that involve participants (Trotter, 1999; Pahl and Spencer, 2004; Emmel, 2008; Edwards, 2010). Social network analysis has attracted increasing attention from by scholars across different disciplines ranging from sociology, business, mathematics and economics (Freeman, 2004). This is because it is an analytical tool that can be used to understand the structural patterns of network of relationships and because networks have to receive increasing attention from international entrepreneurship scholars. 159

179 Most social sciences scholars tend to promote the use of both quantitative and qualitative approaches or methods rather than using one of them in isolation. This helps the researcher to confirm and validate the data and at the same time reap the benefits of using each approach (Creswell, 2003). This tendency is beginning to be focused on combining both types of methods of SNA, especially in the business and management literature which has been dominated by quantitative methods for decades (Monsted, 1995; Coviello, 2005; Edwards, 2010). Social network analysts argue that using SNA not only offers the benefit of making the best use of both methods but it also represents a unique platform to do so (Coviello, 2005; Edwards, 2010). In other words, because of the unique nature of social networks that encompass both structural and process dimensions, using quantitative methods can help to understand the different structural properties of the networks ("outsider view"). In addition, using qualitative techniques help us to understand the underlying processes or content of the networks through gaining "an insider view" of network (Edwards, 2010:5) 24. There have been calls to integrate qualitative methods to SNA, not to undermine the value of quantitative techniques but to serve as a complementary method to enhance the understanding of underlining the network processes (Crossley 2009; Emmel and Clark, 2009; Heath, Fuller and Johnston, 2009; Edwards, 2010). On the one hand, quantitative methods of SNA can help in mapping networks of different kinds and calculating structural measures such as the size of network, centrality measures and density of relationships -that helps in gaining an external view of network (Edwards, 2010; Coveillo, 2005). On the other hand, qualitative methods of SNA help in understanding and addressing the content of relationships and context in which 24 Social network analysis has its ancestry in mathematics of socio-metry, graph theory and ethnographic research. In addition, anthropologists used SNA to study the structure of affinity and interpersonal relationships. SNA therefore has its roots in both quantitative and qualitative approaches (Barnes 1954; Edwards, 2010). However, SNA started to increase in popularity in the last four decades when the development of user friendly statistical packages was produced such as Pajek and UCINET software (De Nooy et al. 2005; Borgatti, Everett and Freeman 2002). This software helps in handling network data and facilitates its analysis. 160

180 networks are embedded. Furthermore, they focus on the understanding the nature of interaction between actors, strength and value of relationships. Moreover, traditional quantitative methods map the network at a specific point of time (snapshot) and fail to capture the dynamic and evolutionary nature of networks. An illustration of limitation approach is given in Figure 4.3. Mapping of network ties involve converting them into binary data (i.e. the presence or absence of tie). This is a common feature of mapping and dealing with network data (Hanneman and Riddle, 2005). Also, to fully understand the strength/value of a relationship it is not enough to record the frequency of interaction (Edwards, 2010). Some actors may interact frequently more than others because of the nature of work they do. That does not necessarily mean that a strong tie exists between them based on trust. For example, a client and a supplier - located in different places- contact each other on a regular basis in order to create a prototype of an integral component to the client company. The presence of regular interaction does not mean that breach of trust cannot occur. It is only through positive interaction and experiences that trust is established, especially in the absence of a formal contract. Thus, in order to better understand the nature of interactions between individual actors, and the differences, replicability, and dynamics of networks, it is essential to use both qualitative and quantitative methods to map and measure structural aspects of networks (Edwards, 2010). Figure 4.3: A traditional quantitative method and its limitations It is equally important to distinguish between data gathered from the whole network and from parts of a network, as each approach could serve a different research objective. Defining network boundaries and sampling It is a challenge to set the boundaries of a network as they may be composed of a large number of actors depending on the population of interest. Therefore, it is useful to think of different strategies to collect data about relationships among different actors in the form of a 161

181 spectrum. At one end of the spectrum, lies the "whole network /full network approach" which has to do with gathering data from all actors of a particular network. This tends to be costly and time consuming and in some cases it is impossible to gather the required data (Hanneman and Riddle, 2005). At the other end of the spectrum, a sample of the whole network is taken, which focuses only on a subset of relations. This approach is less costly and conclusions could be generalized to some population (Hanneman and Riddle, 2005). The decision of which route to follow depends on the research questions and objectives. These two alternatives and the possibilities they offer are summarised in Table 4.7. Table 4.7: Comparison between different sampling approaches Approach Definition Advantages and Disadvantages Whole /Full Data is collected from each It helps the researcher network actor's ties with all other gather a full picture of the actors. In essence, this structure and properties of approach is taking a census network. of ties in a population of actors. In other words, data However, it is costly to are collected about all ties collect data and difficult that are present between all to implement. dyads in a population. Best used in It is very useful especially if the analysis involves a small group of actors. Snow ball approach The researcher often starts off with a focal actor(s) and then they are asked to name other actors they relate to. Those other actors then nominate others. This continues until some type of closure could be achieved, such as no new names are mentioned. This approach is very effective to capture the elite network data. Members who are not connected to this network (often referred to as isolates) are not included. It tends to overstate the "connectedness" and "solidarity" of populations of actors. In addition, it is not always possible to find all of the connected individuals in the population. Very useful for particular types of populations. For example, members of an elite club or kinship networks. 162

182 Ego-centric networks (with alter connections) The researcher starts with a set of focal nodes (egos) and identifies the actors (alters with whom they are connected). After that those alters ego is connected to are asked to identify how they are connected to each other. Alternatively the ego may be asked how his alters are connected to one another. This is due to the fact that the point of start could be with a really marginal/wrong place leading to loss of whole sub-sets of actors who are connected - but not attached to starting points. It is quite effective for collecting a form of relational data from very large populations; it can give a good and reliable picture of the kinds of networks (or at least the local neighbourhoods) in which individuals are embedded. Also, it enables the researcher to understand the opportunities and constraints that ego has as a result of the way they are embedded in their networks. Useful if whole network data cannot be gathered or gathering data about an ego s local neighbourhoods is important for the research. Allows only the gathering of micro-data and measures about network rather than overall network measures. Ego-centric networks (ego only) This involves gathering data about the ego s connections to alter(s) without their connection with one another. It is helpful if the objective is to identify which actors have many friends and which have only a few. This enables the researcher to understand the variation in actors places in social structure, and make some predictions about how these locations constrain their behaviour. It is useful if the objective is to gather data about social roles played by an ego and local social structure in which ego is embedded. It is a good approach to understand the actors in terms of their social roles rather than as individual occupants of social roles. Accordingly, information about local social 163

183 structures could still be identified. The nature of the whole network cannot be identified using this approach. *Adapted from Hanneman and Riddle, 2005 In practice, SNA scholars have combined quantitative and qualitative research strategies together through using different research methods to collect and/or analyse the data. The main purpose is to enhance the triangulation of data through using different research methods and/or to provide complementary evidence and explanation of phenomena under investigation (Edwards, 2010; Creswell, 2003). Figure 4.4 illustrates two studies that combined qualitative and quantitative research methods to enhance triangulation and to offer two different interpretations or perspectives on a particular topic. 164

184 Lievrouw et al. s (1987) study on intellectual networks between biomedical scientists; used quantitive methods to produce maps of networks and interviews to understand history of relationships and contextual characteristics of network. In both of her studies, -which focused on the network dynamics of new ventures and types of resources that flows through a NV s network of relationships, Nicole Coviello focused primarily on collecting data through conducting qualitative multiple case studies using in-depth interviews with entrepreneurs/founders of new ventures. After conducting a thematic content analysis of the qualitative interviews, the structural properties of network were computed using UCINET software and a visual map of network were drawn and confirmed by interviewees later on (Coviello, 2005; 2006 Coviello and Cox, 2006). Simultaneously, content analysis helped in better understanding of dynamics of network, perceptions of actors and identification of vital contextual factors that helped in better interpretation of structural properties (Coviello and Cox, 2006;Coviello, 2006). Furthermore, interviews helped the researcher to understand the "life-story" of the network and how it evolved (Coviello, 2005:43). Accordingly, she adopted a research design and method that encompassed both "soft" and "hard" data (Coviello, 2005:40) which she referred to as a "bi-focal approach" to data analysis, that enabled her to understand the structural and dynamic processes of networks (Coviello, 2005: 40). Figure 4.4: Illustrative examples of using mixed methods to enhance triangulation This is in line with Monsted s (1995) argument that to understand the processes of merging networks we need to use both research strategies. He highlights the fact that some types of ties such as "very weak ties or latent ties" are usually not recorded in matrices used in quantitative analysis and can only be captured qualitatively (Monsted 1995: 194, Edwards, 2010). Another concern highlighted by Crossley (2009) is that focusing on the quantitative approach to SNA "oversimplifies" the social world of networks and neither the structural configuration of network nor its measurement can be divorced from its content (p.8). To sum up, combining methods helps the researcher not only to understand the structural properties and mapping of networks but also to understand their context and dynamics 165

185 (Edwards, 2010). It also helps the researcher to have an internal (perception of individual actors) and external view of the network (structural properties and mapping). After providing the background of key research design and methods informing this study, the following section now illustrates how these designs and methods were applied. Research design and methods used in this study This study used multiple case studies to understand the dynamics of Egyptian and UK SMEs networking and internationalisation processes over time (Eisenhardt, 1989; Yin, 2009). In order to understand these evolutionary processes, a retrospective approach was used based on the key respondents memory and recall of each company s life story since it was founded (Bryman and Bell, 2003) 25. Twenty SMEs were studied from each country Egypt and the UK allowing for a sufficient number of cases to identify key patterns and themes (Lamb et al., 2009). These cases were collected from a variety of sources as explained in Table 4.8. Table 4.8: Sources of UK and Egyptian SMEs Cases UK SMEs Egyptian SMEs Sources Thirteen of the companies were either selected from a compiled list from Thomson One Banker database and the Oxford Science Park website or from the researcher s personal contacts. Seven SMEs were identified through the help of the Birmingham and the Black Country Chamber of Commerce. Four SMEs were identified through the researcher s personal contacts. Sixteen SMEs were identified through the help of the Entrepreneurs forum in Alexandria and the Social Fund for Development (SFD) in Egypt. These cases were purposefully selected and the bases of theoretical sampling were firms that had less than 250 employees, have been operating for at least 10 years and have been 25 Ideally a longitudinal approach would have been better for collecting data from multiple case studies. However, O Donnell and Cummins (1999) argue that it is not always feasible to do so, especially in this context where the focus is on well established firms with 10+ years of history. 166

186 engaged in international markets for at least 6 years. This is to allow the researcher to capture the SMEs internationalisation behaviour from conception to stability and maturity. The choice of a cut-off figure of 6 years is assumed to be sufficient for the SME to demonstrate its capacity to survive in foreign markets. Further information on the two national samples is given in Appendix 1.1. The interviewees were the founder(s) of the company or middle level professionals who were responsible for the international development of the company (such as sales, export and business development managers). The unit of analysis in this study is the network from the perspective of the SME s decision-maker (entrepreneur) the so-called ego network perspective (Hanneman and Riddle, 2005). Ego the entrepreneur defines the boundary of the network from his own perspective, and the interviewees were asked to identify key connections between different actors to provide a relatively comprehensive picture of the configuration of the network. Although SMEs connections in foreign markets could be independent, possible connection(s) between actors were probed for. For example, if the entrepreneur mentioned a key relationship in foreign market who is an old school friend and previously he mentioned another friend who helped him to internationalise in domestic market, I asked about the possible connection between these two friends. This is similar to the ego-centric network analysis with alter connections. Moreover, the possible connections between ego s alters is important as it could offer an understanding of the degree of cohesiveness (density) of ego s network. This is a measure of proportion of ego s alters that are connected to one another (Coviello, 2006, Hanneman and Riddle, 1998). The different network structural measures will be discussed in Chapter

187 A total of 60 SMEs were contacted in the two countries out of which 42 SMEs agreed to participate in the research (twenty from the UK and twenty-two from Egypt). Two were excluded from the Egyptian sample because these were SMEs that had either just started exporting and/or focused only on one market and a few one-off transactions. The UK SMEs were contacted initially by describing the purpose of the study and a request for participation, followed by a follow-up call to remind the respondents about the and to negotiate access. In the Egyptian context -where use of is not a norm, telephone conversations were the initial point of contact and several follow-up calls were carried out to confirm the dates and access. This resulted in a total of 52 semi-structured interviews in the two countries that were conducted by the researcher through site visits during the period 2009-early Each interview lasted on average between one and half to two hours and produced pages of single spaced transcripts. Fifty interviews were tape recorded and transcribed and only two cases refused to permit tape recording. In addition, extensive field notes were taken by the researcher to document the interview. Multiple informants were interviewed whenever possible to confirm/refute data reported by other informants (Johanson and Vahlne, 2003). In seven of the UK SMEs, multiple informants were interviewed (either co-founder(s) and/or export manager) and in Egypt only four cases, multiple informants were interviewed 27. Questions within the interviews depended mainly on open-ended questions, as this is believed to provide deeper insights and explanation by each respondent in his /her own terms. The use 26 All the Egyptian interviews were conducted before the start of the 25 th of January 2011 revolution. 27 Egyptian SMEs are usually managed and owned by one person. Export/business development managers were hardly found in SMEs. 168

188 of open-ended questions suffers from some difficulties as it requires more effort and time to collect and code different responses. However, this type of questioning in a qualitative study is deemed appropriate as it does not "require specific answers as compared to quantitative study" (Bryman and Bell, 2003:37). In order to increase the validity of measures, interview questions were piloted among two practitioners and one academic to avoid any leading and ambiguous questions. The interview covered key themes namely (1) background information about the company (2) concept generation and establishment history (3) internationalisation pattern, history and outcomes (4) key contacts and relationships that were significant for the firm s growth (internationalisation) and survival over time (Coviello, 2005: 2006; Coviello and Munro, 1997; Yli-Renko et al., 2002; Pratshantham and Dhanraj, 2010). Regarding each network attachment the respondents highlighted, I used probing questions to identify the tie s history and initiation, the pattern of tie development, reasons behind such development and key benefits for internationalisation the SME derived from this tie. Finally, the strength of tie was measured within the course of the interview through the use of concentric circles approach commonly used by sociologists (Borgotti et al., 2000). The aim is to identify varying degrees of strength for each group or type of actors. This meant that after asking about the full history of SME from inception until the time of data collection, I grouped key ties mentioned by the interviewee into categories and asked him/her to locate the strongest tie to the SME on the closest circle to the centre (SME) and weakest tie on the farthest circle and any medium strength relationships to be located in middle circle. However, in some cases the interviewee identified the strength of a particular tie(s), within the course of our conversation. This was taken into account as a key representation of the strength of that particular tie. 169

189 In addition, four interviews were carried out with trade advisors and managers of governmental supporting agencies in the UK and Egypt in order to understand the type of support these supporting agencies offers to SMEs and criteria through which SMEs become eligible for different types of support. Additional secondary sources were also gathered including brochures, reports and drafts of export promotion laws that were enacted in recent years 28. Copies of the interview schedules used to interview both the SMEs and supporting agencies presented in Appendix 1.2 and 1.3. Interview transcripts collected from the Egyptian SMEs were translated from Arabic to English by the researcher. To check the quality of my translation, three transcripts were back- translated from English to Arabic by a professional bilingual translator/editor. Data arrangement and preparation After transcribing the interviews, I did not immediately code the interviews as is the norm with qualitative research but rather I wrote the full story of each case and the chronology of events (including the timing of each market entry and tie creation) and key triggering events/junctures were identified 29. This was done on an iterative basis, where the case report was revisited several times to ensure the accuracy of chronology of events and identification of key network actors. Each case was treated as an independent test that either confirmed or refuted the emerging themes (Brown and Eisenhardt, 1997). 28 I was given a copy of the Egyptian export promotion law from the CEO of Social Fund for Development which is the main governmental supporting agency in Egypt. 29 These historical chronologies are useful milestones for identification of patterns of development over time (Pettigrew, Woodman, and Cameron, 2001). 170

190 I then prepared a large spreadsheet where I identified key events and presented them in the corresponding chronological order, where each event was represented by a row. An event could be an internal affair of the company, entry to a new market, further penetration of existing market, identification and creation of a tie or termination of a relationship. For each row I classified the ties involved into several categories as shown in the following example (see Table 4.9). 171

191 Table 4.9: Individual tie coding based on different tie characteristics the Manu5 case. Codes Description Illustrative Example Event Tie Basis of tie initiation Strength tie of Nature of tie An event could be for example, an internal affair of the company, entry to a new market, further penetration of existing market, identification and creation of a ties or termination of a relationship. Identification of key ties involved in particular event This could be a tie that is intentionally created or reactivated by the entrepreneur (intentionally managed tie). It could be a result of chance or unplanned incident (coincidental tie) and finally it could be a tie that was a result of a third party referral (semi-coincidental tie) Could be strong, medium, weak strength based on concentric circles and the way the interviewee described a certain tie. Could be a business, social or a mixed relationship. Whereas a business relationship is mainly contractual or work related relationship. A social relationship is mainly personal relationships with family and friends. Finally a mixed relationship is a multiplex relationship that has both business and social (personal) dimensions. Manu5 s founder approached Mr H. - who is an old school friend - to refer him to his relative in USA. Mr. H and Mr H s relative Relationship with Mr H was intentionally managed. However, relationship with Mr H s relative is semi-coincidental. Relationship with Mr H (as interviewee mentioned) is strong; and his relative is a weak tie. Relationship with Mr H is a social relationship and with the relative is a business relationship The interviewee was asked how the relationship with Mr H s relative evolved. He mentioned that was a medium term relationship that last only until the project was finished and he said it was not a friendship relationship. 172

192 Tie perpetuation Key benefits driven from tie Relationshiprelated costs Could be a one-off relationship (i.e. one transaction) or a medium term relationship (for example, a tie that lasted only for a certain period of time and ceased to exist). For example, SME might have a relationship with a particular project manager and this relationship ceases to exist after the project finishes. Finally a continues relationship, is a relationship that is on-going, Could be a strategic referral, market knowledge and entry to market. For example, lack of commitment, breach of trust. Medium, lasted only until the overseas project was over. Mr H served as a bridging relationship and his relative was the medium through which the company entered the US market. Company had to find another agent (other than Mr H s relative) to operate in US market. 173

193 As described in Chapter 3, in order to capture the dynamics of social capital and SME s internationalisation process, three stages of firm life cycle were proposed as shown in Figure Pre-Internationalisation stage Involves concept generation, legal establishment of the company and the first commercial activity. Early internationalisation stage Involves the first/concurrent international activities and is characterized by limited focus on overseas markets. Subsequent internationalisation stage Involves further internationalisation of the firm and is usually characterised by extensive international involvement and perhaps the use of more advanced modes of entry (such as: joint ventures). This is usually an intentional decision taken by the SME decision maker to aggressively internationalise. This could be a proactive decision or as a result of an external stimulus (i.e. reactive decision). Figure 4.5: SME s internationalisation stages The boundaries of these stages are not rigid but rather they are evolutionary and process based. The movement from one stage to another is identified by trigger events or juncture points in the firm s lifecycle. 31 As will become evident in Chapter 5 and 6, the data obtained indicated the validity of applying this stage categorisation to all the SMEs studied although it was originally proposed in the new venture model (Figure 3.1). 174

194 Data analysis Based on a chronology of events, three adjacency matrices were prepared for each case in a separate spread sheet and UCINET 6 and Netdraw were used to map and measure different structural dimensions of the SMEs network 32. Each case had three different matrices reflecting the SMEs network in each stage, and the tie pattern of development was taken into account 33. After preparing the data and calculating the network structural measures for each of the three stages of SMEs life cycle, I followed a mixed approach for analysing data, which Coviello (2005) referred to as a " bi-focal approach " (p.40). This mixed approach involved two steps. The first step involved preparation of a summary spread sheet which included key internationalisation performance measures over time and the structural measures of SMEs network for both UK and Egypt. I then was able to conduct some non-parametric statistical tests (such as Spearman correlations and Mann Whitney U test and Chi-square tests) to identify internationalisation and networking evolutionary and coevolutionary patterns. The next step involved a quantitative analysis of the data through using the exploratory hierarchical clustering technique to obtain three main taxonomies of UK and Egyptian SMEs respectively. 32 These matrices are binary and directed, which shows whether a tie exist or not and its direction. Following that an attribute file for each stage per case was prepared in order to visualise the different types or categories of networks present within each stage of the SMEs life cycle. Drawing a clear distinction between individual and inter-organisational networks is multifaceted (Mitchell 1969; Coviello, 2005; Larson and Starr, 1993). The focus of this study is on the relationships between individuals and between these individuals and different organisations. 33 UCINET 6 is a social network analysis software that allows researchers to calculate different structural measures for overall network of SME and individual egonets (networks of each actor within a given network) (Borgottii et al, 2002). NETdraw is a software that converts these matrices to a network map so that different actors within each stage of firm development is mapped and drawn. 175

195 I then began to identify key themes and differences between internationalisation and networking processes of SMEs within each cluster within each case and across cases. This was done on an iterative basis until I reached a level of theoretical saturation where no more themes of differences could be identified (Lamb et al., 2009; Yin, 2009). To ensure the accuracy and reliability of data, I draw on different sources of evidence through using of multiple informants and secondary sources such as minutes of board of directors, periodic reports, annual reports, ongoing list of action plans, websites and promotional material. The use of documents is argued to serve as a significant complementary source of evidence which might either confirm or contradict interview data; this enhances the triangulation of data (Yin, 1989; 2009) 34. The details on how data were triangulated are provided in Table Table 4.10: Triangulation of research data Interviewee(s) Secondary sources How triangulation was applied UK Mainly were founder(s); export and business development managers There were mainly project databases; company s website; newsletters; promotional materials (brochures) and general company profiles and annual reports. As a general practice I consulted the company s websites before contacting the company and kept a record of their nature of business and latest news they published on their website which helped me to have a general overview of the company s activities and to probe. For example, Egypt Mainly were founder(s) and on three occasions (production/sales managers; head of publication department). Internal s, management meeting records and company detailed profiles and financial statements. Project databases and descriptions and promotional material. Similar to the UK whenever possible I compared the documented material about projects and action plans to the interview transcripts but in four cases I had an access to internal s, management meeting records and company 34 In the two of the Egyptian cases, documentary evidence on dates of new market entry were at variance with the information supplied in the interviews which led me to re-contact the interviewees to check further. 176

196 I would start by saying "I have seen on your website that you operate in so and so markets and can you please take each market individually and talk about the story behind entering this market?" I then asked at the end of the interview if they had a documented record of projects/clients (in mentioned markets) with names and dates and details. I was given this list whenever it was available. Hence when conducting data analysis I compared the interview transcripts with the projects/clients details and annua reports to confirm the dates and details. detailed profiles and financial statements. Also, from these documents I double checked and identified some of the ties or relationships (especially with agents and overseas clients) that were mentioned in the interview. Moreover, I provided feedback covering some of the key findings to two of the UK SMEs and three of the Egyptian SMEs to ensure that I have collected data that matched their experiences. I also cross checked with another researcher the illustrative quotes reported in Chapter 7 and there was over 90% agreement regarding the way I interpreted these quotes 35. Summary of the research process The research process followed in this research is reflexive and iterative. I started with reviewing the general literature on internationalisation, networking, institutional theory and co-evolution. Based on which two co-evolutionary models were developed for two types of SMEs, namely traditional SMEs and new ventures (see Chapter 3). Moreover, the review of literature guided the data collection and preliminary analysis. 35 This inter-rater reliability was only used for the antecedents section mainly because this is the only part where different interpretations could emerge. As for the identification of chronology and history of the SME, it is a straight forward process that requires less interpretation. 177

197 After completing data collection and the first stage of data analysis, it became apparent that there was a mismatch between the proposed theoretical frameworks and the empirical findings. SMEs tended to choose different internationalisation and network development patterns at different points in time. SMEs hardly followed a uniform or linear internationalisation pattern and they were likely to be shaped by critical incidences emerging from the external environment. This called for a refinement of the proposed stages of internationalisation discussed in Chapter 3. In order to answer the first three research questions, I carried out three rounds of data analysis. Firstly, I prepared a chronology of events for each case and identified the structured patterns of internationalisation and network development. This was done through codification of the structural aspects of network and the internationalisation performance outcomes over the three internationalisation stages of the SME. Moreover, I wanted to check if there are coevolutionary trajectories between the two processes. I relied on social network analysis software (UCINET6) to compute the structural dimension of network and Spearman correlations to find out if there are patterns of association (co-evolution) between structural dimension of networks and internationalisation performance measures over time. This analysis illustrated the general tendencies of the Egyptian and UK SMEs. Secondly, to understand the co-evolutionary trajectories within each sample, I carried out an exploratory clustering of SMEs. This clustering resulted in six taxonomies of UK and Egyptian SMEs. It served as a way to find out key similarities and differences between the different clusters. I then began analysing the accounts provided by the SME decision makers, to understand how and why these co-evolutionary trajectories took place over time. The qualitative findings offered in-depth explanations as to why and how internationalisation and network processes co-evolved and how the SME responded and sometimes influenced changes in the external environment. These findings are presented in Chapter

198 Thirdly, the qualitative findings provided important explanations regarding how certain contingencies (such as the national institutional context) shape the SME s behaviour, as shown in Chapter 7. Guided by on the empirical findings discussed in Chapters 5-7, I shall propose a revised co-evolutionary framework and compare it to the two a priori frameworks discussed in Chapter 3, (please refer to Figures on pages ). Based on the previous discussion, it is difficult to associate myself with a particular theoretical orientation. In other words, I am neither a pure positivist nor a pure interpretivist. Also, I do not hold the view that reality is socially constructed by the social actor or that reality is completely independent of the researcher. I fit somewhere between the two extremes. I am not a subjectivist or an objectivist. I am a critical realist [CR]. Critical realism does not render itself to a particular ontological or epistemological consideration, but rather calls for the use of abductive reasoning to understand the social phenomenon (Welch et al, 2011). Furthermore, Welch et al. (2011) suggest that understanding why a certain outcome/effect took place can only be interpreted through identifying the causes and origins of this outcome. They argue that social scientists need to dig beyond the observable facts to understand why actors behave in a certain way or why a certain process took place. This calls for the abolition of the trade-off between inductive theory building and deductive theory testing dichotomy. In addition, qualitative research methods allow for an understanding of how processes and causal mechanisms develop over time and provide explanations to why certain causalities and processes took place through the interpretations of experiences provided by the social actors. For their part, quantitative research methods are crucial for "calibrating the understanding of a certain phenomena and in uncovering the prevalence or predominance of actors experience" (Lee, 1999; Bluhm, Harman, Lee and Mitchell, 2010:5). 179

199 This calls for a mixed methods approach to complement different aspects of the phenomenon under investigation (Blum et al., 2010). In other words, quantitative methods help to test and identify causal relationships usually deduced from established theories. While qualitative research can offer contextualised and novel explanations of causal mechanisms identified through a detailed investigation of individual(s) experiences (Welch et al., 2013). To meet my research objectives I used comparative multiple case studies. I collected data from twenty UK and twenty Egyptian cases. I relied on semi-structured interviews with key SME decision makers and gatekeepers in the governmental supporting agencies in both the UK and Egypt. I relied on mixed methods to analyse data or in Coviello s (2005) terms a bifocal approach. Quantitative methods were used to pinpoint internationalisation and network development patterns and to test for presence of co-evolutionary trajectories. On the other hand, qualitative methods were used to provide a contextualised explanation of the internationalisation and network co-evolution. Conclusion This chapter has explained the philosophical orientation guiding this research. It also has provided a detailed explanation of multiple case study design used in this research and how the qualitative data was collected and analysed. I also explained the unique approach through which the findings of this study were analysed. The next chapter provides details on how the key measures used in this study are operationalised and then the analyses of the co-evolution of SME internationalisation and network development. 180

200 Chapter 5 OVERALL CO-EVOLUTIONARY PATTERNS: NETWORKS AND INTERNATIONALISATION DEVELOPMENT PATTERNS Introduction This chapter presents findings on the co-evolutionary patterns of SMEs social capital development and internationalisation processes. The choice of SMEs operating in foreign markets for at least 6 years allows their development from conception to maturity to be examined. As mentioned in chapter 4, these cases were purposefully selected, to provide differences in firm age, geographical coverage, international sales revenues and mode of foreign market entry. Following Prashantham and Dhanraj (2010), variations in the firm s initial stock of social capital was allowed and subsequent variations in the dynamic social capital were also permissible (Saxienian, 1994; 2002). This chapter is to provide an overview of networking development patterns and the internationalisation development over time; and the co-evolution of the two processes, network development and internationalisation, for both UK and Egyptian based SMEs. The following section describes how these findings were prepared and presented. 181

201 Generic framework and data preparation Given the complexity of the data and number of variables involved, a generic framework was prepared to help the reader to understand the logic and discussion of different network and internationalisation outcomes of the SMEs. Figure 5.1 presents the guiding framework used in this chapter. 182

202 TE- t2 TE- t1 Figure 5.1: Generic Framework t 0 t n t 1 t 2 X 1 X 2 X 3 Network Variables: Network Variables: Network Variables: Network Variables: Net size T 1 Net Size T 2 Net Size T 3 Average net size across time Net Density T 2 Net Density T 3 Average net density across time Net Constraint T 2 Net Constraint T 3 Average.net constraint across time Co-evolution line Internationalisation outcomes variables: PFST 2 Internationalisation outcomes variables: Internationalisation outcomes variables: PFST nno. of Markets T n Time line Time period/stage No. of Markets T 2 No. of regions T 2 PFST 3 No. of Markets T 3 No. of regions T 3 No. of regions T n Rt.Tn Legend: X 1 :Pre-internationalisation stage(pis) which involves concept generation, legal establishment of the company and the first commercial activity. X 2 : Early internationalisation phase (EIS)- which involves the first/concurrent international activities and is characterized by limited focus on overseas markets X 3 : Subsequent internationalisation phase (SIS)- which involves further internationalisation of the firm and is usually characterised by extensive international involvement (which is usually a response to a trigger event or sometimes just the proactiveness of the entrepreneur) T 0: Timing of pre-internationalisation phase T 1 : Timing of early internationalisation phase T 2 : Timing of subsequent internationalisation phase T n : Present time Te-t1: Trigger event at time T 1 Te-t2: Trigger event at time T 2 Net size: Network size at T 1 ; T 2; T 3 Net Density: Network density at T 1; T 2; T 3 Net constraint: Network Constraint at T 1; T 2; T 3 Av. net size: Average network size across time Av.net density: Average network density across time Av.net constraint: Average network constraint across time PFS: Percentage of foreign sales at T 2; T 3; T n No. of markets: Number of markets at T 2; T 3; T n No. of regions: Number of regions at T 2; T 3; T n Rt : Rate of growth at T n 183

203 As shown in Figure 5.1, SMEs pass by three main phases through which there is assumed to be interplay between network variables (structural variables) and the internationalisation outcomes/performance. On the assumption that these two processes co-evolve together and influence one another, it is imperative to highlight the evolutionary patterns within each of the three stages and also across the three stages (i.e. over time). This was done by going through the qualitative interview(s) of each individual case and formulating a chronology of events for each case. This involved identifying key relationships that were at each of the three stages of SMEs lifecycle, identification of history and development of these ties and the internationalisation history within each stage. In addition, key trigger events or conjunctions that had a significant impact on the way the firm operates were taken into account and accordingly cut-off dates through which the company moved from one stage to another were determined. This enabled me to draw a time line for each case and to identify key network relationships and internationalisation outcomes that took place within each stage and helped to depict the networking and internationalisation behaviour of the SMEs over time. Having this time line to hand allowed me to take a longitudinal view on the evolution of networking and internationalisation behaviour of SMEs and also to have a retrospective view (at the present time t n ) regarding the present status quo of SMEs network and internationalisation performance. The logic behind this guiding framework is similar to the "net present value" 36 concept in finance and life insurance fields, whereby the company s (life cycle) starts at time t0 or time zero and then as the time moves on, the company passes through different time periods 36 Lin, Grier and Nagalingam (2000). CIM justification and optimisation. London: Taylor and Francis. p

204 (t 1 and t 2 ) and then the lifecycle stops at time tn which is the present time 37. This allows the researcher to examine what happens within each time period and also to have an assessment of the company s current status quo at present time. Accordingly, three network variables; namely network size, network density and constraint were calculated for the pre-internationalisation stage [PIS], the early internationalisation stage [EIS] and the subsequent internationalisation stage [SIS]. Also, three internationalisation outcomes were calculated for the two stages of actual internationalisation namely EIS and SIS. These network and internationalisation variables/outcomes are referred to as "within stage" measures. Furthermore, to gain a dynamic view of the network variables the averages of the network size, density and constraint in each stage were calculated. Also, internationalisation outcomes over time were computed. Table 5.1 describes in detail the meanings and operationalization of the key variables used in this chapter. 37 Present time refers to year

205 Table 5.1: Definitions and operationalisation of key variables Key Measures Within stages Network structural measures Variables Description Comments Code/ Network size Network density Refers to the number of ties or contacts in ego's network, in other words how many contacts or ties ego is connected to, this is a proxy to the degree to which ego is involved in networking (Arenius,2002; Coviello, 2006; Borgatti, Evert, Johnson, 2013). In this research network size refers to the number of ties in a n SME's network at a given period of time. It is a measure of network cohesion and it is the ratio of number of ties in a given network to the possible number of ties. It is also the probability that a certain tie might exist between any two actors in a given network (Borgatti et al, 1998; 2013; Coviello, 2006). In this research network density refers to the degree of connectedness of SME's contacts at a given period of time and a density score of 1.00 or 100% means a completely dense network where everyone knows everyone (star network). As the SMEs network size increases this strictly speaking means that ego has more access to resources and opportunities. This increases the SMEs social capital and creates a degree of self-sufficiency, whereby the company becomes less dependent on a limited number of contacts. (Narooz and Child, 2012; Arenius, 2002). Coleman (1988) and Burt (1992) have two divergent views regarding the impact of dense networks on ego's performance. On one hand, Coleman (1988) argues that highly dense networks have a closure effect whereby most of the actors in ego's network are connected and these ties are characterized by a high degree of trust and commitment, actors have reciprocal interests. These strong closely connected ties serve as key sources of rich information. On the other hand Burt (1992), argues that the higher the density of a given ego's network the more likely it will have redundant ties and repetitive sources of information and ideas. Hence the higher the density of ego's network the less social capital s/he has. However, if ego has access to structural hole (Burt, 2005) - becomes connected through a bridging relationship to a network s/he has no access to - s/he is more likely to have access to new sources of information and opportunities and becomes more innovative (i.e. he has more social capital, than another ego who does not have access to structural hole or to the broker(s). Closely linked is the fact that ego himself could be a broker and connect otherwise unconnected networks and hence becomes more powerful (Burt, 2000; 1998; Coviello, 2006; Narooz and Child, 2012). abbreviation Net Size Net density Density measures should be interpreted carefully as small network is more likely to be dense (in a particular context) than a large network (in a different context). 186

206 Borgatti et al. (2013) gave an example of a small department in business school; a density score of 0.5 might be small as most of the staff members are likely to know one another. However, in another context like an entrepreneur s network with a density score of 0.5 might be high because his network might be composed of people inside and outside his own home country. Therefore it is better to use density when comparing between two networks and its interpretation is highly dependent on the study context (Borgatti et al., 2013). 187

207 Within stages Network structural measures (Cont'd) Network constraint This is a measure of structural hole whereby if ego connects two unconnected alters he or she is a broker (where s/he can manipulate the network for his own benefit) and if he is connected to a bridging relationship that connects two unconnected alters ego then he has access to structural hole through his connection with bridging relationship (Burt, 1995; 1998; 2000; 2005; Borgatti et al., 2013). Hence, network constraint is a measure of the ''extent to which ego's alters have ties to each other and it is an indication to the extent to which ego invests time and energy on alters who invest in each other'' within a given period of time (Borgatti et al., 2013:275). Ego's investment in an alter is measured by the amount of ties this alter has (Borgatti et al, 2013). A network with a large number of structural holes signifies higher level of social capital where SME has more diverse access to resources and diverse sources of information. As there are fewer redundant ties (which provides redundant information). So if the SME has a high level of network constraint this means that the network has a limited access to structural holes (Burt, 2005;1998) and the SME is said to be obscured by alters in his or her network(burt,1992; Borgatti et al., 1998; 2013). Net constraint Within stages International isation outcome measuresthese are international isation performance outcomes Percentage of foreign sales to total sales (FS) Number of markets Number of regions Refers to percentage of foreign sales to total sales at a given period of time, given that some SMEs have both inward and outward internationalisation. The respondents were asked to combine both sales revenue coming from outward and inward and estimate the proportion of these two out of the total sales. Refers to the number of overseas markets (countries) SME operates in at a given period of time. Refers to number of regions in which SMEs operate in namely (Asia; Europe; North America; South America; Africa; Australia; Middle East) at a given period of time. If an SME s foreign sales are more than the local or domestic sales, this is an indication of high international involvement and higher the scale (depth)of internationalisation. The higher the number of markets an SME operates in, the greater the scope of SMEs internationalisation. The greater the number of regions, the higher the breadth (i.e. international diversity) of SMEs internationalisation. PFS No. of markets No. of regions 188

208 Within stages Network structural measures (Cont'd) Network constraint This is a measure of structural hole whereby if ego connects two unconnected alters he or she is a broker (where s/he can manipulate the network for his own benefit) and if he is connected to a bridging relationship that connects two unconnected alters ego then he has access to structural hole through his connection with bridging relationship (Burt, 1995; 1998; 2000; 2005; Borgatti et al., 2013). Hence, network constraint is a measure of the ''extent to which ego's alters have ties to each other and it is an indication to the extent to which ego invests time and energy on alters who invest in each other'' within a given period of time (Borgatti et al., 2013:275). Ego's investment in an alter is measured by the amount of ties the alter has (Borgatti et al, 2013). A network with a large number of structural holes signifies higher level of social capital where SME has more diverse access to resources and diverse sources of information. As there are fewer redundant ties (which provides redundant information). So if the SME has a high level of network constraint this means that the network has a limited access to structural holes (Burt, 2005;1998) and the SME is said to be obscured by alters in his or her network(burt,1992; Borgatti et al., 1998; 2013). Net constraint Within stages International isation outcome measuresthese are international isation performance outcomes Percentage of foreign sales to total sales (PFS) Number of markets Number of regions Refers to percentage of foreign sales to total sales at a given period of time, given that some SMEs have both inward and outward internationalisation. The respondents were asked to combine both sales revenue coming from outward and inward and estimate the proportion of these two out of the total sales. Refers to the number of overseas markets (countries) SME operates in at a given period of time. Refers to number of regions in which SMEs operate in namely (Asia; Europe; North America; South America; Africa; Australia; Middle East) at a given period of time. If SMEs foreign sales are more than the local or domestic sales, this is an indication of high international involvement and the higher scale (depth) of internationalisation. The higher the number of markets an SME operates in, the greater the scope of SMEs internationalisation. The greater the number of regions, the higher the breadth (international diversity) of SMEs internationalisation. PFS No. of markets No. of regions 189

209 Table 5.1: Definitions and operationalisation of key variables (Cont'd) Key Measures Variables Description Comments Across stages network measures Average network size across stages Average network density across stages It refers to the sum of the network sizes divided by the number of time stages/phases. Refers to the summation of network density divided by the number of stages/phases. The higher the average network size of SME, the more it has access to resources and information that are crucial for its survival and growth within domestic markets and overseas. The higher the average network density of SME the more likely that network is highly cohesive and is composed of close strong trustworthy relationships where rich and tacit knowledge is exchanged. However, there is a closure risk whereby SME is restricted and has limited exposure to new sources of information and ideas (Burt, 1992; 1998). Code/ Abbreviation Av. Net size across stages Av. Net Density across stages Average network constraint across stages Refers to the summation of network constraint divided by the number of phases/stages. The higher the average constraint of SMEs network, the more likely that the company has limited access to new sources of information and knowledge (i.e. Limited access to structural hole) and is more likely to be constraint by other actors in its network (Burt, 1998; Borgatti et al., 2013). Av. Net constraint across stages Across stages internationalisation outcome measures Number of markets at t n This refers to an over time view where at time tn, what is the number of overseas markets the SME operates in? The greater the number of markets an SME operates in, the greater is the scope of internationalisation. No. of market t n Number of regions at t n Percentage of foreign sales at t n Rate of growth at t n This refers to a over time view where at time t n, what is the number of regions the SME operates in? This refers to a over time view where at time tn, what is the percentage foreign sales to total sales (at 2012). This is the percentage of foreign sales at time t n divided by number of years being international The greater the number of regions an SME operates in, the greater is its breadth (international diversity) of internationalisation. The higher the percentage of foreign sales of SME, the greater is the scale (depth) of internationalisation. The higher the rate of growth over time, the better the international growth of the SME. No. of regions t n PFStn Rttn Years being international Refers to the number of years SME has been operating overseas until t n. The greater the number of years being international, the higher the SME internationalisation learning and experience. Years being international 190

210 As discussed in the Chapter 2, co-evolution implies that key processes within an organisation coevolve with one another and with the external environment. The SME might respond to a certain stimulus (conjunction or trigger event) emerging from the external environment by adapting to changes in the environment or changing the conditions inherent in the external environment in a way that helps it to grow and survive. The details regarding co-evolution and how companies adapt to change of external environment and how the network development process and internationalisation co-evolve together over time are discussed in more detail in Chapter 6. The objective of this chapter is to provide a holistic view of how SMEs networking and internationalisation behaviour evolve and co-evolve over time. Internationalisation and networking behaviours of SMEs evolve individually and co-evolve over time. It is imperative to distinguish between causation and co-evolution. Internationalisation and networking processes of SMEs co-influence one another, in one occasion SMEs networks have an impact or influence on internationalisation of the SME (for example, a SME might decide to enter a particular overseas market through piggybacking on an MNC and /or the SME might be part of a global supply chain and follow the global buyer all over the world). On the other hand, internationalisation might influence the networking behaviour of an SME, for instance, the SME might decide to enter a certain overseas market to increase its profitability and it will then try to find a suitable agent in that particular market to serve the customers). Thus, it is not possible to position network or internationalisation variables as being dependent or independent in this case The key premise of co-evolutionary perspective used in this research is that; there is no clear cause and effect relationship between the internationalisation and networking processes, but rather the these two processes co-evolve and co-influence one another at different points in time. For example, internationalisation could be triggered by a certain relationship (such as: unsolicited order from an overseas client) at a given point in time and the SME decision maker could decide to enter into a new overseas market at another point in time. Or the SME decision maker might decide to further penetrate a certain overseas market through searching for a new agent and at the same time s/he might also respond to an solicited order in the same market at the same point in time. This means that these two processes influence and cause one another at different points in time (or simultaneously) and co-influence one another in a cyclical manner. 191

211 Hence, both processes are not mutually exclusive they co-influence and depend on one another. In order to test this and to gain some insight on the different patterns of association between networks and internationalisation, a set of non-parametric correlations was conducted to test the correlation between; different network variables within each stage, different internationalisation outcomes within each stage and network variables and internationalisation outcomes within each stage. Finally, a set of correlations was also carried out between different network variables and internationalisation outcomes over time which is referred to as across stages correlations. Since this research investigates the differences between SMEs behaviour in Egypt and the UK, I did not standardise which network variables and internationalisation outcomes to assess for each of the two samples. Instead, a correlation analysis was carried out for all of the above mentioned variables with the aim of identifying the significant correlations between the variables for each sample separately. This might shed some light on the differences in networking and internationalisation behaviours of UK and Egyptian SMEs. Within stage and across stage correlations for both UK and Egyptian SMEs are now discussed. For the breakdown of the measures for each case in the two samples please refer to Appendix

212 A. UK SMEs evolutionary and co-evolutionary patterns To understand the evolutionary patterns of both networking and internationalisation behaviours of UK SMEs, the mean values of each variable and outcomes were calculated across the 20 cases and a simple graphical presentation is provided to highlight the development patterns of networking and internationalisation behaviours. Furthermore, a set of non-parametric (spearman) correlations were conducted within stages and across stages. A.1 Within stage correlations and overall development patterns I shall start by presenting network development patterns over time based on the mean value of the three network variables, followed by within stage non-parametric network correlations. A.1.1 Network development patterns and within stage network correlations Table 5.2 presents the means of the British SMEs network density, size and constraint level within each stage of the SME s lifecycle. 193

213 Table 5.2: Average network measures within stages (UK SMEs, N=20). Time Network size Network density Network constraint t t t It is clear that mean network size increased over time as the company moved from pre-internationalisation stage to subsequent internationalisation stage. However, network density and constraint decreased over time, which indicates an inverse relationship between the network size and network density and constraint. This is depicted in Figure

214 This means that over time UK-based SMEs tend to focus more on networking (building new contacts) and rely on sparse networks (less dense networks characterised by a relatively high level access of structural holes). The tendency to build new contacts and rely on relationships to have access to resources and information was highlighted by several respondents. Manu-12 s export manager mentioned that relations can be viewed as long-term investment: "Something I found very important personally in dealing with relationships is that today s tea boy might become tomorrow s manager..[ ] our guys are always attentive to every one of our clients [..].. Because you never know even the tea boy may not become a manager but he may pass a message to somebody [..].. This (word of mouth referral) costs you nothing it s a very cheap commodity and very under used but not by ourselves we know the significance of word of mouth". 195

215 Also, Manu-3 s founder argued that he would do anything to find the easiest and cheapest way to find a suitable relationship in overseas markets and the possibility that this new relationship might serve as a bridging relationship or a platform to access new sources of information and resources. "I stand to lose everything and also stand to gain everything. We give the best personal service to our clients and suppliers, if someone needs me in Germany next week, I will find the easiest, cheapest and quickest way to go there I recently gave a presentation in front of companies which wanted to export to Germany.and my presentation was on how to do business in Germany and I said the first thing I did I went to three websites Easy Jet, Ryanair and BMI websites and I said where can I fly to, what is the cheapest option and once I have realised that, I went to Hamburg and tried to find a customer there in my level; it doesn t matter if you lose some pennies and hasn t come off that s ok but at least you have [..]. it is just simple things.[ ].. Moreover, relationships are crucial for our business and getting to find out that people can put you in touch is something amazing. [ ] making yourself available for that and that s the only thing that works". This company in particular relied extensively on the bridging relationships it created in individual overseas markets to get access to other key relationships in those markets and even in other countries to gain access to resources, network of distributors, suppliers and referrals to original manufacturers (OEMs). Moreover, some entrepreneurs valued and utilised bridging relationships differently. Some relied on them mainly to have access to information, support or to enter new markets. Others created links with key brokers to reach otherwise unreachable relationships. The following two examples illustrate how SMEs might rely on structural holes differently over time. The first one is OG2 196

216 company which is an oil and gas company and the second one is a traditional manufacturer of multi-lock products. The founder of OG-2 focused on a niche in the market and decided to establish his own company and focus on this niche. He was a former student of a well-known UK-based university and thought of capitalising on his strong relationship with the university to grow and expand his company. He utilised his geographical proximity to large oil and gas companies to access their overseas networks and enter new markets. Also, through referrals from these big MNCs other marine companies began approaching the company (these MNCs served as a platform or a bridging relationship). The links with marine companies enabled the company was able to sell its products through the overseas networks of these marine companies. However, as the company reached the SIS it began to aggressively pursue other markets (in which it has limited or no operations) and build relationships with agents and distributors in these new markets. Another example, Manu-12 (which is a multi-lock manufacturer) followed a different networking approach in the pre-internationalisation stage. The company directly approached local clients who bought anything that could be made out of crude aluminium especially letter plates. They didn t have any intention of exporting until some Scandinavian clients (who were using same type of letter plates like the UK) approached the company and began buying some of its products to be used in their respective markets. Over time the letter plates business died and the managing director decided that it was time to expand and specialise in multi-lock business. The sales manager/export manager then decided that in order to succeed the company had to compete with German manufacturers who were the key players in this industry and he began approaching the distributors and suppliers of OEMs in the UK and overseas in order to get access to OEMs and he managed to convince them that they offered the same good quality of German lock manufacturers but at a lower price. This was a huge success and it even forced the German manufacturers (through their subsidiaries in the UK) to approach Manu-12 and buy from it. 197

217 As the company entered the SIS, it began to think of exporting as the primary means of survival since the profits from home market were diminishing. They decided to de-internationalise for a while from Middle East and Far East countries and focus only on nearby markets such as Eastern Europe. It then approached its local chamber of commerce to help them get access to eastern European markets. Over time the company began to re-internationalise into diverse markets through direct search or reliance on key clients to get access to distributors or suppliers. In one instance the company relied on one of its clients that buys raw materials from China and asked him if he could refer them to someone who could help in finding reliable suppliers. The client then referred Manu-12 to VC company which is a private company that specialises in helping overseas companies to find good, reliable Chinese suppliers. Also, Figure 5.2 infers that the higher the network density, the higher the network constraint (Burt, 1998; 2010; Hanneman and Riddle, 1998; Borgatti et al., 1998). However, if the network becomes too large it becomes difficult to manage and maintain (Coviello, 2006). Also, the larger the number of strong/close relationships (dense network) an ego has the more constraint s/he is because of most of his time and effort are devoted to maintaining a large number of strong relationships, giving him no access to structural holes (Burt, 1992; 2010). Manu-9 is an example of a company that started off with a highly dense network and evolved into a large sparse network which became difficult to manage and also they relied on bridging relationships in subsequent internationalisation stages. Manu-9 is a concrete and construction materials manufacturer. It started as a department in XY Company and in 1982 the current founder joined XY Company as an engineer in this department. Later it became a business unit and began to introduce a new range of metal coatings and 198

218 concrete products. The founder then became the export manager of the SBU. During the preinternationalisation stage, the network density was 100% whereby the network size was only 3 (founder, XY co. and colleagues) with a network density score of 100%- because everyone in ego s network is connected to the other and a network constraint level of which is more than 1.00 (this is considered very high). Within less than one year after the company s establishment the founder was asked by his supervisor to start and target overseas markets to sell the new range of products through forming networks of distributors in each new market. He started by choosing two distributors one in Belgium, whose responsibility was to form a network of distributors and sell to the western European markets, the other distributor in the UAE, who found other distributors in Qatar and Saudi Arabia. Shortly after the success of these exporting attempts, the founder was then asked to establish a network of distributors to target new markets in the world. The founder then approached the local chamber of commerce and UKTI and sought their help. The UKTI team referred the founder to the overseas high commissioners and UKTI teams whereby they offered the founder with a list of potential clients. The founder also did desk research complying names of likely clients or distributors in a particular market and obtained their contact details from telephone directories and began contacting them. The UK embassy in the Czech Republic provided the founder with a list of potential distributors out of which the founder chose one. This relationship evolved into a strong business relationship whereby the Czech distributor referred the company to his twin brother who lives in Sweden and he became the company s distributor. Things followed a similar pattern whereby the founder approached different overseas consulates and acquired a list of potential distributors and approached them or he did his own research selected the ones that proved to be suitable. There were few 199

219 incidences of referrals from clients of distributors. On one or two occasions a few overseas agents approached the company as it began to have a well-known name a key characteristic of the relationship with UKTI is that Manu-9 became a preferential tie whereby it were always given a unique support in overseas markets and advice because of its success and strong relationship it developed with the UKTI and local chamber of commerce. The network size during EIS was 43 and it had a density score of 44%. Moreover, the network constraint level decreased to 0.038, which means that the company relied on bridging relationships to a relatively high extent. Over time the domestic market became more price-driven which made it nearly impossible for the company to stay competitive. Also, the original owner of the XY company started hiring more employees hoping to enhance the sales and marketing efforts. This attempt further increased the cost for the company and the original owner decided to sell the company. Manu-9 s founder and his colleague (a technical director) made an offer to the original owner of XY Company to buy out the division and they registered the company under the name of Manu-9. Both had a strong determination and vision to aggressively internationalise into overseas markets. During the subsequent internationalisation stage the company began to reactivate most of its old ties and relationships with clients and distributors and capitalize on them to enter new markets or obtain referrals to prospective agents. Moreover, several overseas clients approached the company through the website making the website one of the key means of further internationalisation of Manu-9. Furthermore, as Manu-9 s network began to expand it became really difficult to keep track of every single tie and to keep up to date with current development in overseas markets. He relied on existing strong relationships he had with overseas clients and distributors to help company to expand further. 200

220 The founder summarised his network behaviour by saying" when I look at all of these markets we want to enter, relationships are absolutely fundamental from all of these businesses overseas. We bought, we opened a new business [ ] I do not have the time available to meet all of our export customers and distributors I am actually relying in many cases on very strong relationships which I built up and worked hard over more than twenty years I am relying on these relationships now to carry us forward". As Manu-9 s founder noted, network relationships are fundamental and crucial for his business however, if the network became very large over time it would became very difficult to manage and the only way out might be to rely on strong direct and/or bridging relationships he established over time. The same networking logic was mentioned by founder of Con-2. He noted that: "Finding the right people to deal with and seek advice from is essential to our business.. I view relationships as our key business model.[..]..however, it is only me who manages the company and I can t keep in touch frequently with all of our key contacts so I tend to evaluate regularly those relationships that yield best benefits and discard the ones that are not useful or beneficial". Con-2 follows a different rationale when it comes to dealing with large networks; the company tends to follow a calculative approach whereby they maintain those ties that generate benefits and discard those that are not beneficial anymore. These examples raises the question of whether there is a significant correlation between network size and network constraint and density as the SME moves from pre-internationalisation to 201

221 subsequent internationalisation stage. To answer this question a set of non-parametric correlations 39 was carried out to highlight significant correlations (if any) between these three variables over time. These correlations are presented in Table All of the variables were tested using Shapiro and Wilk normality test, Skewness and Kurtosis z-scores to identify which ones approached normal distribution. Most of the variables were not normally distrusted apart from network constraint at time t1 and average density over time and rate of growth at time tn. This is an expected conclusion since both samples were only 20 cases and this is a very small sample. Moreover, the basic assumption of this research is that network and internationalisation variables correlated and each influence one another. So it is not possible in this case to identify which ones are the dependant and/or the independent variable. Hence, non-parametric correlations were used (Shapiro and Wilk, 1965;Razali and Wah,2011;Cramer,1998;Cramer and Howitt,2004; Doane and Seward, 2011). 202

222 Table 5.3: Within stage network correlations (Spearman's rho correlation coefficient) Network density t 1 Network density t 2 Network density t 3 Net size t 1 Net size t 2 Net size t 3 Net constraint t 1 Net constraint t 2 Net constraint t 3 Network density t 1 Network density t 2 Network density t Net size t Net size t Net size t Net constraint t 1 Net constraint t ** ** ** Net constraint t ** * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). *** N=

223 As shown in the Table 5.3, there is a significant correlation between different network variables during the course of the SME internationalisation. During pre-internationalisation stage, none of the network variables showed significant correlation with one another. However, as the SMEs began to internationalise network constraint and network density showed significant positive correlations during the early and subsequent internationalisation stages. This means that as the SMEs network density decreases less cohesive, sparser networks the SMEs level of network constraint decreases; which means that there are more structural holes in the SME s network. Furthermore, there is a negative correlation between network size and network constraint in the preinternationalisation and early internationalisation stages, whereby, as SME s network size increases the SME, tends to depend more on bridging relationships (access to structural holes) as it moves from pre-internationalisation stage [ PIS] to early internationalisation stage [EIS]. This means that the SME builds new relationships as it moves from the pre-internationalisation to early internationalisation stage and relies more on access to structural holes to gain access to new sources of information and resources. Surprisingly, there is no significant correlation between network size and network constraint as the SME moves into the subsequent internationalisation stage [SIS]. Referring to Table 5.2, it is clear that the percentage change in network size is very high as the SME moves from pre-internationalisation stage to early internationalisation stage (53%) which is greater than it is between the early internationalisation and subsequent internationalisation stage (28%). The percentage of change in level of network constraint is relatively high (decreased from 67% to 48%). This may explain why there is a significant negative correlation between network size and network constraint during PIS and EIS. Another plausible explanation could be as the SME moves into subsequent internationalisation stage [SIS] it has enhanced its networking capability and improved their internationalisation experience. This means that SMEs become more active in building new relationships (intentionally managing ties) by themselves rather than relying on referrals or intermediaries (brokers) to do so. Moreover, some of the UK SMEs establish a strong legitimacy and become well recognised in local and overseas markets which encourage clients and suppliers to approach them directly. Also, recently the internet has played a crucial role in the further 204

224 internationalisation of UK-based SMEs. This suggests that access to structural holes is not as important in the SIS as it is at the PIS and EIS. Moreover, at the SIS, SMEs would have advanced their networking capability and gained more power or status (Burt and Merluzzi, 2013) and they then themselves become brokers and start to dictate their own terms on new ties they create. Manu-5 for example, is knowledge -intensive company whereby during the SIS stage the founder managed to create relationship with a large MNC called U co. This MNC has a large global network of distributors/agents in most of the countries around the world. The founder managed to persuade the MNC to let the SME become its UK-based agent. In the international sales meeting the founder met up with Mr P who is the international sales director of U co. and they got along pretty well. Mr P then Left U co. and opened his own business called P company in Mexico. The founder then approached Mr P and asked him to be his agent in Mexico and that Manu-5 would be the company s agent in the UK. Both Mr P and Manu-5 founder formed a mutual agency agreement capitalising on each other network of distributors. Moreover, they also utilised the connection they had with U company s network of distributors worldwide and through this network they would sell their products around the world. Also during the SIS through the profound and established profile the company has, a large distribution company in India approached the company through their website and suggested becoming their agent and supplying their products to Asia. The founder wanted this distributor to form and find agents in Australia and New Zealand and supply the company s products through them. This shows how the company s enhanced network capabilities and status helped it to negotiate favourable terms with MNCs and a large distribution companies around the world. 205

225 A.1.2 Internationalisation development patterns and within stage internationalisation correlations The SME is likely to pass through two internationalisation stages mainly EIS (t2) and SIS (t3). The means of the two internationalisation outcomes are summarised in Table 5.4 and Figure 5.3. Table 5.4: Internationalisation outcomes within stages (UK sample, N=20) Time Percentage of foreign sales No. of foreign markets No. of regions t 2 32% 7 4 t 3 53%

226 Figure 5.3: within stages internationalisation development patterns The percentage of foreign sales to total sales during the EIS and the SIS has increased from 32% to 53% which is an indication of increased the scale of British SMEs internationalisation over time. Furthermore, breadth and scope of internationalisation is measured by number of regions and overseas markets in which SME operate has increased from the EIS to the SIS. The number of foreign markets served increased from 7 to 19 and the number of regions increased from 4 to 5. This demonstrates that UK-based SMEs increased the depth and the breadth of their internationalisation as they became more engaged into overseas markets and moved to the SIS. The internationalisation rationale and behaviour differ as the company becomes more accustomed to internationalisation and gains more confidence in its ability to excel in overseas markets. The founder of Manu-3 stated that: 207

227 "When we get the first inquiry and the prices are low we don t want to lose any money, but we are not greedy so once we get the order we fulfil it and so forth for two times and then we started to increase the price. Because it s important to get into new markets and in many different markets and seven- eight years ago we really had nothing". Also Manu-9 s founder argued that: "Probably in about mid 1990 s (start of subsequent internationalisation stage).we were looking at a map of the world and we had a map on the wall and started asking ourselves where in the world we are not operating or not doing any business. What are we targeting?. I m sorry to say but it s really hard work and I carry on talking about kissing frogs if you know the old story of kissing frogs occasionally, bingo you get the conversion you want.you have to be very enthusiastic and keep going.". In order, to check if there is a significant correlation between the three internationalisation outcomes within stages, a set of non-parametric correlations was conducted and the results are presented in Table

228 Table 5.5: Within stage internationalisation correlations (Spearman's rho correlation coefficient) Percentage of foreign sales t 2 Percentage of foreign sales t 3 Number of regions t 2 Number of regions t 3 Number of markets t 2 Number of markets t 3 Percentage of foreign sales t 2 Percentage of foreign sales t 3 Number of regions t 2 Number of regions t 3 Number of markets t ** * Number of markets t ** * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). *** N=20. As shown in Table 5.5, there is a significant positive correlation between number of markets and number of regions in the early and subsequent internationalisation stages respectively. Moreover, there is a significant correlation between the number of regions in t h e early internationalisation stage [EIS] and number of regions at the subsequent internationalisation stage [SIS]. This suggests that internationalisation breadth increases over time in terms of number of markets and number of regions in which SMEs operates and that British SMEs tend to internationalise and expand to a largely geographically dispersed markets during EIS and SIS respectively. Furthermore, it could be argued that there is a likelihood that if the SME starts to pursue diverse and largely geographical dispersed markets in the EIS, they are more likely to continue to do so in the SIS. 209

229 A.1.3 Within stages co-evolutionary correlations As shown in the previous examples, SMEs rely on network ties to have access to needed resources and information and to cease new opportunities as they emerge. It is crucial to see which of the network variables have a significant correlation with internationalisation and which ones do not. Table 5.6 illustrate the correlation between network variables and internationalisation outcomes within each stage. Also, the possibility that one variable might be correlated with another one in a subsequent stage is also investigated. For instance, it is worthwhile investigating if network constraint in EIS has an influence on the number of markets in which a company operates in SIS, because if there is a negative correlation between these two variables one could argue that the more the company relies on referrals or bridging relationships in the EIS stage, it is more likely to be more geographically diverse in subsequent internationalisation stage because of the opportunities that were made readily available for the company through those bridging relationships. 210

230 Table 5.6: Within stage co-evolutionary correlations (Spearman's rho correlation coefficient) Net. density t 1 Net. density t 2 Net. density t 3 Net size t 1 Net size t 2 Net size t 3 Net const. t 1 Net const. t 2 Net const. t 3 Percentage of foreign sales t 2 Percentage of foreign sales t 3 Number of regions t 2 Number of regions t 3 Number of markets t 2 Number of markets t 3 Network density t 1 Network density t 2 Network density t Net size t Net size t Net size t Net constraint t ** Net constraint t ** ** Net constraint t 3 Percentage of foreign sales t **

231 Table 5.6: Within stage co-evolutionary correlations (Spearman's rho correlation coefficient)"cont'd" Net. density t 1 Net. density t 2 Net. density t 3 Net size t 1 Net size t 2 Net size t 3 Net constrai nt t 1 Net constrai nt t 2 Net constrai nt t 3 Percentage of foreign sales t 2 Percentage of foreign sales t 3 Number of regions t 2 Number of regions t 3 Number of markets t 2 Number of markets t 3 Percentage of foreign sales t 3 Number of regions t 2 Number of regions t 3 Number of markets t 2 Number of markets t ** ** ** ** * * ** * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). *** N=

232 As shown in the above correlation matrix, there is a positive correlation between network size and number of regions and number of markets at EIS; and a positive correlation between network size and number of markets in SIS. This strongly suggests that the SME s network size increases, the SME s internationalisation breadth increases (in terms of increased number of markets and regions) 6. Surprisingly, there is no significant correlation between the network size and percentage of foreign sales. This suggests that British SMEs rely on their existing network contacts to diversify and increase the number of their overseas markets across different regions. There is a significant negative correlation between network constraint at the PIS stage and number of markets in SIS. This implies that British SMEs rely more on intermediaries and/or third party referrals during their early internationalisation stage in order to gain access to information and resources that they would later utilise to enter new markets during subsequent internationalisation stage. A.2 Across stages co-evolutionary correlations We have seen that network variables correlate with internationalisation outcomes within the early and subsequent internationalisation stages and they might even influence the evolution from one stage to another. For example, the level of network constraint in PIS was highly negatively correlated with the number of markets in SIS. This suggests an accumulation of knowledge and experience from third party referrals that would later facilitate the SMEs entry to new markets in the subsequent internationalisation stage It is important to note that the opposite process also applies, whereby as the SME enters new markets it is more likely to form new relationships in those markets. 213

233 It is worthwhile investigating if these arguments hold true over time; taking into account the current status quo of UK-based SMEs. Therefore a set of measures were calculated to assess if any of the network variables influence and are influenced by internationalisation outcomes over time. These are measures that take time into account and draw our attention to the status of SMEs at the present time. Table 5.7 demonstrates the co-evolutionary correlations over time. For example, the percentage of foreign sales at time t n is the percentage of foreign sales that was reported by the SMEs for the period between 2010 and The number of markets and regions is the total number of markets and regions the SME operates in without taking into account the chronology of events. As for the network variables, the average network size, average network density and average network constraint were calculated for each of the three stages The average network size is equal to total network size of the three stages divided by the number of stages. The same logic applies to network density and constraint. The researcher computed the accumulated figures of network density, size and constraint to test if different correlations with internationalisation outcomes would appear. There was no significant differences between accumulated and average figure of the three network variables would yield a different significant correlation with internationalisation. The reason behind this is that network measures computed for each stage takes into account the ties that were established during this stage. 214

234 Table 5.7: Co-evolutionary correlations across stages (Spearman's rho correlation coefficient) Average net density over time Average network size over time Average net constraint over time No. of years being inter. Percentage foreign sales t n Number of regions t n Number of markets t n Rate of growth t n Average net density over time Average network size over time Average net constraint over time No. of years being inter Percentage foreign sales t n Number of regions t n * Number of markets t n * ** ** Rate of growth t n ** ** * *Correlation is significant at the 0.05 level (2-tailed). **Correlation is significant at the 0.01 level (2-tailed). *** N=20 As shown in Table 5.7, there is no significant correlation between any of the network variables across stages of internationalisation. In other words, average network size, average network density and average network constraint are not correlated as is the case within stages. There is a significant positive correlation between the average network size and number of markets at Tn. This suggests that over a period of time, British SMEs rely extensively on relationships to enter new overseas markets. This is illustrated by the founder and managing director of Lek Company: 215

235 "Our networking strategy is that of a sales and engineering oriented company. So the position is that sales have to come and services have to be provided and engineers fix them and then we get paid. So we do anything in our capacity through being sales aggressive finding relationships needed to do that whether inside or outside the company. We have for each sales person and managers a target. Also, each one of the manager sums every night saying how good or bad he/she is doing and we make a competition between our different branches.each branch manager gets an saying to him what other branches say in Poland or Portugal has done (in terms of maintaining crucial relationships and finding new markets),and only the managers can see it. Moreover, the co-founder of OG1 company argued that without referrals and word of mouth the company would have not been able to become a global company (this company operates in 6 continents and in 14 markets). He said: "Our internationalisation was a slow build up process over a period of time. As people got to know us, friends and J (co-founder) has a very good personality and gets along with nearly everyone. What helped us to internationalise globally is the word of mouth. We have tried to advertise using brochures.we have also tried exhibitions and we found that we are spending thousands and thousands and we could not really find the return from it So relationships with large MNCs (oil and gas companies) has helped us in our internationalisation spread and kept us moving forward until this date. Furthermore, there are some interesting correlations between different internationalisation outcomes over time. The number of years being international is positively correlated with number of regions and number of overseas markets. However, it is negatively correlated with 216

236 the rate of growth at time tn. This implies that as the SMEs learning and international experience increases, they gain more confidence to internationalise into a larger number of diverse markets. There is a negative correlation between the number of years being international and rate of growth 42, the overall average rate of growth of the British SMEs is 4% and the average number of years the UK-based SMEs spend operating in international markets is 24 years. One plausible explanation is that the longer a SME spend in operating in overseas markets the lower the yearly growth rate. Although there are some companies in the sample which have been operating in international markets for over 30 years, the objective here is just to give an overview of the internationalisation behaviour of the SMEs. Individual cases will be discussed more closely in the next chapter. Moreover, the percentage of foreign sales is positively correlated with the rate of growth. This suggests that the higher the percentage of foreign sales the higher is the growth rate of the SME. Furthermore, the number of overseas markets is positively correlated with the number of regions and is negatively correlated with the rate of growth 43. This suggests that UK -based SMEs focus more on market diversification and becoming global companies more than they focus on increasing the level of their international sales revenue. This might be attributed to the SMEs piggybacking a multinational that operates on a worldwide scale. SMEs are often part of a global supply chain and this is reflected in a large number of diverse markets that the SME has access to. For example, Manu-1 is a conveyor manufacturer and they tend to be part of a global supply chain of conveyor handling and manufacturing. In its early days the company was the main 43 It is important to note that it is easier for a smaller firm to achieve a high rate of growth (%) than a larger firm (i.e. relative growth) as opposed to growth in absolute terms. 217

237 provider of conveyors used by the British royal mail. Since the UK has the oldest Royal Mail system in the world, they were responsible in helping other countries in installing and operating mailing systems. This involved having British expatriates working with people in overseas markets and most of these expatriates know about Manu-1 through the former interaction (as Manu-1 was the main conveyor supplier of Royal Mail) and they asked the company to supply the overseas mail and post offices. During the subsequent internationalisation stage the company was one of the key airport conveyor manufacturers and suppliers in the world and they tended to supply most of the airports around the world. Another example, OG1 is an oil and gas company and it sells safety signs for marine and oil and gas companies overseas. They started their early internationalisation through Shell (large MNC) and through its links with other overseas oil and gas companies; OG1 was the main supplier of these overseas oil and gas companies in different oil and gas capitals around the world. B. Egyptian SMEs evolutionary and co-evolutionary patterns To understand the evolutionary patterns of both networking and internationalisation behaviours of Egyptian SMEs, the mean values of each variable and performances/outcomes were calculated across the 20 cases and a simple graphical presentation is provided to highlight their patterns of development at the three points in time. Furthermore, a set of non-parametric (Spearman) correlations were computed both within stages and across stages. In other words, the same analysis was followed as for the UK SMEs. 218

238 B.1. within stage correlations and overall development patterns I shall start with presenting network development patterns over time based on the mean value of the three network variables, followed by within stage non-parametric network correlations. B.1.1 Network development patterns and within stage network correlations Table 5.8 gives the means of the Egyptian SMEs network density, size and constraint level within each stage of the SME s lifecycle. Table 5.8: Average/mean values of network variables within stages (Egyptian SMEs, N=20). Time Network size Network density Network constraint t t t In general, network size increased from the pre-internationalisation stage [PIS] to the early internationalisation stage [EIS]. However, it remained the same as SMEs moved from early internationalisation stage to subsequent internationalisation stage [SIS]. Network density and network constraint decreased over time as the SME moved from the PIS to the SIS. The evolution of the three network variables is presented in Figure

239 Figure 5.4: within stage network development patterns There is an inverse relationship between network size, network density and network constraint. Also, there is a positive relationship between network density and network constraint within each of the three stages of SMEs life cycle 44. To evaluate the significance of these correlation and whether this hold true within each stage of SME s life cycle, a set of Spearman non-parametric correlations was conducted to test the relationship between variables. These are presented in Table Out of the 20 Egyptian SMEs, 6 companies did not reach the subsequent internationalisation stage. This leaves 14 companies which reached the subsequent internationalisation stage. For the sake of simplicity those companies that haven t reached the third stage were given a score of zero. This might deflate or decrease the average figures used Table 5.8 and figure 5.4 Therefore, it is important to be aware of this aspect when interpreting the results. By contrast, only two of the UK cases reached the subsequent internationalisation stage. 220

240 Table 5.9: Within stage network correlations (Spearman's rho correlation coefficient) Network density t 1 Network density t 2 Network density t 3 Network density t Network density t Network density t * Net size t ** Net size t 1 Net size t 2 Net size t 3 Net size t * Net size t * * Net constraint t 1 Net constraint t 2 Net constraint t ** ** Net constraint t ** * ** Net constraint t 3 Net constraint t ** * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). *** N=

241 The correlation matrix presented in Table 5.9, shows that network size is negatively correlated with network density and network constraint during the PIS. This implies that the smaller the size of a SME s network during the PIS stage, the more dense the network (cohesive) and it is likely to have limited access to new sources of information and opportunities (embedded in networks with several bridging relationships that offer access to structural holes). As shown in Table 5.8, the mean of network size during PIS is 10, which indicates that Egyptian SMEs have a relatively small network during the pre-internationalisation stage and a relatively dense network with a mean score of 30%. Network density in the pre-internationalisation stage is positively correlated with network constraint. This implies that the denser an SME s network during the PIS, the less likely it will depend on weak bridging relationships to gain access to new sources of information and opportunities. Egyptian entrepreneurs tend to rely on their close friends and colleagues to seek support and help. If close friends in entrepreneur s network cannot help, they would then ask them for personal recommendations to help solve a particular problem or to obtain the information they need. In other words, it is the immediate close ties that help entrepreneurs have access to certain brokers. For instance, FR-1 is an example of a company that relied heavily on the founder s network of close strong relationships. Over time, it created other strong and close relationships and devoted every possible effort was devoted to keep the dense network intact. FR-1 s founder- Mr M -graduated from one of well-known public universities in Egypt and during his university years he established a portfolio of close friends. After graduation each of those friends went on his own way and when the founder decided to open a company, the 222

242 founder approached two of his friends as he thought they would surely help him. They provided him with the capital he needed to start the business and left its management fully in his hands. Then one of his friends Mr S who became a leather manufacturer referred the company to one of the large public department stores in Egypt (R stores). Through them the company began to expand and grow locally. In addition, this leather manufacturer referred the founder to other suppliers of wood that were reliable and cheap. The sales and production managers of the same company established strong relationships with top and lower level managers of the Omar Affandi department store (the largest public store in Egypt) and they directly approached the store. They said that establishing relationships with lower level managers not just top managers, was crucial because these lower level managers serve as key sources of insider information and are a very good source of referral. For example, they tend to let the company know if there is a project coming up (say an order coming from a resort in Sharm El Shiekh) and ask the company to prepare a proposed design of furniture which they will then show to their top level managers, as a result, the company gets a preferential chance to be selected. These managers also referred clients to FR-1 if they requested something that the stores did not offer. This helped the company to expand into 42 different governorates and cities around Egypt through the network of branches this store had. This shows that dense networks whether those of the founder or established during PIS are key to the company s survival and growth, especially within the local market. Because of these friends help and the dense network, the company was able to gain access to sources of support from the most successful export 223

243 promotion agency, called IMC or expo-link (these agencies offer the most reliable source of support if you manage to get into them through being connected through a key player in the field or they pick successful companies). Fr-1 managed to export for the first time through an American friend of the founder and then the founder was approached by IMC and through the IMC the company participated in several fairs whereby local and international clients bought and made orders from the company. Since then this has been one of the key ways through which the company got into overseas markets. However, building strong relationships with overseas clients took time and because of the poor reputation of Egyptian products they have to see the product first and this might be one of the main reasons that companies focus more on the density of networks than on their size. In a similar fashion, during the early internationalisation stage [EIS] network density and network constraint are positively correlated and both were negatively correlated with network size. By contrast, network density and size are positively correlated in the subsequent internationalisation stage [SIS]. This suggests that if SME s network size increases, that network becomes more cohesive and the SME relies on a limited number of bridging relationships (since there is a positive correlation between network density and network constraint in SIS). Interestingly, a network variable in one stage could be correlated with another network variable in a subsequent stage. For instance, network constraint in EIS is positively correlated with network density in SIS. This means that the less the SME relies on bridging relationships (i.e. access to structural holes) in the early internationalisation stage, the more likely it will have a denser (more cohesive) network in the subsequent internationalisation stage. One explanation could be that the company during the EIS relied on bridging relationships to reach out for contacts with which they have no direct relationship, but over time the bridging relationships 224

244 decayed and those ties, that were disconnected from ego during EIS, become stronger and closer relationships over time due to frequent interaction (Burt, 2001). The tendency of Egyptian entrepreneurs to value strong close relationships and do everything they can to maintain their closeness, is consistent with the collectivist culture of Egyptian entrepreneurs. In addition, network density in the EIS is positively correlated with network size in the SIS. This suggests that the more dense (cohesive) the SME s network in the early internationalisation stage the larger the network size will be in the subsequent internationalisation stage and vice versa. As highlighted by some of the cases, costs of internationalisation increased over time as SMEs became more internationally involved. This may have forced some entrepreneurs to rely heavily on the dense strong networks and they had in the PIS and the EIS to gain access and referrals to key gatekeepers in domestic and overseas markets to mitigate against the costs of increased internationalisation. For instance, FR-2 (a manufacturer of hand -made furniture) depended heavily on family and friends who immigrated in US and Canada to help him expand in the USA during the EIS, and over time with increased competition and key internal conflicts that arose from different US agents. The founder capitalised heavily on close friends and family members in the US to sort out the problems with US clients and to help him find other prospective agents with their own network of distributors. This suggests that as the Egyptian SMEs become more involved in overseas markets the more likely they will rely on Egyptian immigrants in overseas markets to offer support and also to refer them to new contacts that might help the companies to further penetrate overseas markets. Another example is EB s founder, who is a doctor (Dr AS). He graduated from medical school and had a colleague (Dr O) from university whom he approached and told him about his idea, and they agreed to work together (both became the co-founders). Dr AS capitalised on his 225

245 relationship with a former school friend with whom he had lived in Kuwait and studied in the same high school. This friend is of a Syrian origin and he helped the founder to find critical ingredients he needed to manufacture the product. Also the friend came and taught the company the required know-how at a relatively cheap price. The other co-founder used to be a medical representative of a Saudi company in Egypt with a large number of distributors and distribution companies. The co-founder depended on his strong relationship with these distributors across the three stages of the firms lifecycle. In the beginning they helped the company to distribute the product in the domestic market and to establish a strong presence in the home market. The first co-founder also capitalised on personal relationship to know about what support is offered for SMEs and he managed to acquire a very useful piece of advice about the Social Fund for Development and that they offer loans for SMEs and tax exemptions for beneficiaries up to ten years. During the EIS stage they relied on the relationship the second co-founder had with his former employer and the Saudi distribution companies he dealt with to promote the product and sell it to Saudi Arabia. The first co-founder referred him to a friend who knows someone in one of the biggest distribution companies in Yemen, and through this link he managed to enter Yemeni markets. After some time this company stopped exporting and placed the business on hold, and both cofounders went to university and managed to get an MBA. This was to help them learn how to manage the company. Through their links with university professors they were given important advice on management and finance. Through a mutual friend, an export manager was found and appointed for a year to help the company learn how to export. This helped the company to move into the SIS where it reactivated the overseas and local relationships it had placed on hold and asked these ties (especially ones in overseas markets) to refer them to other distributors or 226

246 get referral to enter other markets. They mentioned that relying on friends (informally) helps the company to know information about- for example- where to get something, what support is available to it or who to contact in company X and that these referrals played a significant role in a company s survival and growth and led to an increase in foreign sales revenue and diversity of markets. B.1.2 Internationalisation development patterns and within stage internationalisation correlations Table 5.10 and Figure 5.5 depict the internationalisation patterns of the Egyptian SMEs over time. These SMEs continued to internationalise into the same number of markets and regions as in the early internationalisation and subsequent internationalisation stages. One plausible explanation for this is that Egyptian SMEs continued to focus on further penetration of existing markets and enriching and reinforcing relationships that were developed during the PIS and EIS stages. This is closely linked with the unlimited change in SME network size from EIS and SIS and the limited increase in percentage of foreign sales to total sales. This raises an important question, whether there is a connection between the network size, density and constraint levels and the limited advancement in the Egyptian SMEs internationalisation. One way to answer this is to test if there is a correlation within stages between network and internationalisation variables. Table 5.10: Average/mean values of internationalisation outcomes within stages Time Percentage of foreign sales No. of markets No. of regions t 2 37% 6 3 t 3 42%

247 Figure 5.5: within stage internationalisation development patterns To further understand the internationalisation patterns of Egyptian SMEs across the EIS and SIS stages, a set of non-parametric correlations are presented in Table

248 Table 5.11: Within stage internationalisation correlations: Percentage of foreign sales t 2 Percentage of foreign sales t 3 Number of regions t 2 Number of regions t 3 Number of markets t 2 Number of markets t 3 Percentage of foreign sales t Percentage of foreign sales t Number of regions t Number of regions t ** Number of markets t * ** * Number of markets t ** ** *. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed). *** N=20. As shown in Table 5.11, the number of regions and number of markets during the early internationalisation stage are positively correlated. However, the percentage of foreign sales during EIS is not significantly correlated with any of other internationalisation outcomes. This suggests that the higher the number of overseas markets in which the SME operates the more geographically diverse is the SME s international operations. However, in SIS the three internationalisation outcomes are positively correlated. This suggests that the lower the percentage of SME s foreign sales, the more likely that the SME will operate in a limited number of geographically diverse markets and vice versa. This could be attributed to the fact that Egyptians view exporting to Europe and the western world as a key 229

249 challenge and only few exceptions manage to export to western countries. These were mainly software application providers. For instance, the founder of Fr-3 -which is a furniture manufacturer-, stated that: It is embedded in our culture that if you export to Europe that is exceptional [...]. meaning that it s very difficult and only few ones (companies) do that..[.].so it is easier to sell to Arab countries and the Middle East rather than Europe and USA as they have more stringent requirements and different consumer tastes". On the other hand, other entrepreneurs mentioned that they can do exceptionally well in overseas markets and that they could even excel over foreign competition as the founder of SF-2 argued: "No we can export to any country we want. If they can manufacture useful and important software,we can excel over them and that is what happened with one of our products ;we excelled over the only German manufacturer who manufactures these statistical software and we even won his clients over [ ].. They were mainly European clients. However, I have put off the aggressive penetration into the European market until I have enough cash flow". There are some important correlations between the three internationalisation outcomes across the two stages. The number of markets in EIS is negatively correlated with percentage of foreign sales and number of regions in SIS. This has some implications on the internationalisation development patterns pursed by Egyptian SMEs. It suggests that if the SME operates in a limited number of markets in the early internationalisation stage, it is more likely to have more foreign sales from a diverse number of overseas markets in SIS and vice versa. 230

250 Given the overall stability in the number of markets or regions, different dynamics may be going on which will be discussed in chapter 6. So far, we have seen that the network and internationalisation evolutionary patterns within stage. However, it is worthwhile investigating whether there is a correlation between the network variables and internationalisation outcomes. And there are significant correlations between network variables and internationalisation outcomes are they similar to the UK-based SMEs? B.1.3 Within stage co-evolutionary correlations Table 5.12 shows within stage co-evolutionary correlations. These are correlations between internationalisation and network variables within stages. 231

251 Table 5.12: Within stage co-evolutionary correlations (Spearman's rho correlation coefficient) network density t1 network density t2 Network density t3 net size t1 net size t2 net size t3 net constraint t1 net constraint t2 netconstrt3 percentage of foreign sales t2 Percentage of foreign sales t3 number of regions t2 number of regions t3 Number of markets t2 Number of markets t3 network density t1 network density t2 Network density t * net size t ** net size t * Net size t *.452 * net constraint t1 net constraint t2 Net constrain t3 percentage of foreign sales t2.957 ** ** **.476 * ** **

252 Percentage of foreign sales t3 number of regions t2 number of regions t3 Number of markets t2 Number of markets t3.670 ** ** ** **.631 ** * * * ** * ** * ** ** ** *.600 ** * * * * **.473 * * ** ** * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). *** N=20 233

253 As shown in Table 5.12, the within stages correlations revealed some interesting findings. Network density during the pre-internationalisation stage [PIS] is positively correlated with the percentage of foreign sales, number of markets and number of regions in the subsequent internationalisation stage [SIS]. On the other hand, network size in PIS is negatively correlated with percentage of foreign sales and number of markets in SIS. This implies that a highly cohesive (dense) network composed of strong ties is more crucial for the internationalisation performance of SMEs during the SIS than is the number of contacts that the SME has in the PIS. In addition, the denser the SMEs networks during the PIS the better the internationalisation performance of the SME during later internationalisation stages. This could be attributed to the richness of information provided by the social capital of the founder(s) and the favours that could be sought from close and strong relationships that existed during the PIS. Several Egyptian entrepreneurs said that they do not like to lose relationships but instead focus on nourishing them even if they are no longer immediately useful. They also have the mind-set that some relationships which are not useful today might be beneficial tomorrow. This networking logic was highlighted by some respondents. For instance, EB s co-founder stated that: "Definitely, some of our distribution companies with whom I used to deal with [ ] we used to even meet on daily basis and then I stopped dealing with them because simply the business between us finished [..]... We stayed in touch and we call each other on special occasions (such as special holidays or Eid) and then I find myself at some point that I need something from these companies [..]...a certain service at a certain point in time or something unrelated to their services but someone within that company who knows someone in another company who can help me. So I reactivate this relationship [ ].. However, it s 234

254 very important to note that I never leave a relationship in a disrespectable way so when I need to reactivate this relationship it is not difficult to do so. I never leave a relationship with hard feelings or in a disrespectable way but rather it is our nature. For sure, today we may have mutual benefit but tomorrow we may not, or a contradiction may take place where today we do not need each other but tomorrow we will. So, it is very important to keep the minimum personal aspect of the relationship going". Similarly, the founder of Manu2-eg noted that relationships are crucial and that one could never know who might be useful in the future and who might not be: "Sure, it s a nature of business according to the Egyptian proverb, today s friends are tomorrow enemies and today s enemies could be tomorrow s friends". On the other hand, during the early internationalisation stage network size is positively correlated with the number of markets. Moreover, the level of network constraint is negatively correlated with the number of markets. This suggests that during the early internationalisation stage, Egyptian SMEs focus more on establishing new relationships and rely on access to structural holes to enter new overseas markets (for example, through fairs organised by governmental and non-governmental supporting agencies). This access to structural holes could be referrals from pre-existing relationships or third party introductions. This could be linked to the type of relationships that the SME rely on to enter new markets in the early internationalisation stage. 235

255 Manu3-eg 45 founder argued that: "Those ones (new ties created in early internationalisation stage) were actually mainly personal [ ]..Let me say it s about 60% personal efforts whereby I visited those clients I met in the international trade fair and sometimes even approached some local ones alone and showed them several offers and met them at least 3 times and then they could see that you really are interested and then you might be able to offer good quality products..[.]. As for the international trade fair it helped us with three or four overseas clients because through the three days of the fair, three or four businessmen who came and saw our product(s) thought it is good -because we usually have samples with us and offer good pricesnormally these businessmen rely on big companies such as flora and fine which are very expensive and when they saw our products and good price and quality they decided to buy from...[ ]. So the fair helped us a bit and we also depend on our own personal effort s whereby I visited the client and I usually give them a good offer and good prices we usually introduce the product to them and offer them best prices and offers and this is how we grew". As for the subsequent internationalisation stage, surprisingly, all three network variables (network constraint, network density and network size) are positively correlated with number of markets, number of regions and percentage of foreign sales. This shows that Egyptian SMEs networks are likely to be highly cohesive and composed of some bridging relationships. In other words, they try to achieve a balance between dense (cohesive) 45 Manu3-eg is a paper manufacturer; they produce toilet papers, tissues and paper towels. They have a strong reputation in the Egyptian market and tend to piggyback on large MNCs such as Flora and Fine into Middle Eastern market. However, the company managed to export directly to other Arab and North African and European countries. 236

256 networks and some access to structural holes. This is what Burt (2009:2) referred to as" network duality". This balance enables the firm to enhance its internationalisation performance. An example of network duality is illustrated by the sales director of FR1: "In old days, when there was a common proverb between traders that keeping your word is worth more than million pounds there were no problems.. Now we have more problems although we have contracts but you can always get a clever lawyer to get you out of them or out of the obligations specified in the contract very easily. I usually ask my friends to refer me to a good lawyer.but also having strong personal relationships and trust is of utmost importance for our international growth ". As FR1 sales director emphasised, the company can establish a new relationship for example with a lawyer to sort out problems as they arise, but still they need to depend on their dense and close relationships to internationalise further. In addition, an interesting finding came out of the correlation matrix presented in Table Network density and constraint in the SIS are negatively correlated with the number of overseas markets the SME operates in during the early internationalisation stage. In other words, Egyptian SMEs with limited international coverage in EIS are more likely to rely more on intermediaries and relatively dense networks to further internationalise in SIS. A good example of this implication is Con2-eg, which designs and supervises composting and recycling facilities and manufacture composting products-it operates in Asia and Arab/middle eastern countries. 237

257 The company relies on the dense network of its founders and the linkages they have with other sister companies or affiliated companies to the founder(s) of sister to get access to projects in municipal and agricultural waste composting; and recycling facility design and construction supervision. According to Con2-eg s co-founder: "In the last 15 years (namely during the PIS and EIS) I have moved from being a dancer - excuse me for using this example- you accept what comes in your way, now after years of operations (during SIS).we are well established we pick what only suits us". From the above quote it is apparent that the Con-2-eg relied on the dense network relationships of its three co-founders such as the relationships the first co-founder managed to establish with key ministries in Egypt. Over time the company grew and began to choose which relationships to work with from the pool of clients that approached the company. Nonetheless, they relied to some extent on access to structural holes which the co-founder referred to as a catalyst. "Of course, if you never worked in a particular market, you are like a stranger who is blind although he can see. I gained market knowledge from governments and municipalities not individual companies (company s main overseas clients). So we see who in government of a particular country can help us gain needed information and form an agreement with him or her.[ ]..You often search for what I call a catalyst someone who enhances the process but has nothing to do with management of operations- which is usually between Con-2-eg and the client which is usually a country; who can help us enter and learn about systems, markets and laws in respective markets". 238

258 B.2 Across stages co-evolutionary correlations A key question remains - do these significant co-evolutionary correlations hold true over time? Hence, another set of non-parametric correlations is carried out between the network variables and internationalisation outcomes over time (see Table 5.13). 239

259 Table 5.13: Across stages Co-evolutionary measures (Spearman's rho correlation coefficient) Average net density over time Average network size over time Average net constraint over time No. of years being inter. Percentage foreign sales t n Number of regions t n Number of markets t n Rate of growth t n Average net density over time Average network size over time Average net constraint over time ** No. of years being inter Percentage foreign sales t n ** ** Number of regions t n * Number of markets t n * * ** Rate of growth t n * ** * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). *** N=

260 As shown in Table 5.13, the average network density over time is positively correlated with average network constraint over time. This means that as the SMEs network density increases the network constraint increases (i.e. access to structural holes decreases) In other words, the denser an SME network the less likely it will rely on access to structural holes. The number of markets and percentage of foreign sales over time are positively correlated with average network density and constraint over time, and the average network density is positively correlated with rate of growth over time. This implies that network density and limited reliance on bridging relationships played a more important role in the Egyptian SMEs internationalisation over time than did the number of contacts inherent in the SMEs network. This seems to be a recurring theme across the 20 Egyptian cases. For example, the founder and sales manager of FR1 stated that when it comes to overseas clients close and strong trust worthy relationships are of utmost importance because of the poor reputation of Egyptian product in overseas market. He said: "Sure, product is everything; in Milano s fair it s a proof of what I m saying people there will never deal with an Egyptian just based on a talk or a chat. They have to come and see the product themselves and inspect the quality. [..] After displaying our product at the Milano furniture fair.[ ]. They (overseas clients) dealt with us but for us to go around Europe and tell people I have these products and even show them catalogues they will not deal with me so our overseas business wasn t through just 46 It is important to highlight that usually very small networks show a higher degree of density if most of ego s contacts know one another. 241

261 talking to people but rather good and quality products that helped us gain their trust.. as I said if I go by myself to these people overseas and tell them I have this company in Egypt that does this and this they will not know me. Over time this trust and close relationship is impartial. For instance, now I have a very good relationship with Denis the guy from USA the issue of trust is not a problem we trust each other completely.. So I mean if he comes after two years from now to ask for something I will deal with him without thinking twice I will deal with him as if I m blind. Also, the founder of EB company views overseas relationships as partners rather than mere clients of distributors and he can draw on their support or advice at any time. "Sometimes my own strategy towards a certain relationship changes for example, at some point I focused more on wholesalers but later I became more consumer driven.so I focused more on retailers or end consumers so automatically I ended the relationship with wholesaler-only the business side of it not the personal side. so even the business relationship ended our relationship is very simple I can easily call any of them and ask about something or get certain advice [ ] all of our relationships with agents and distributors is like business partners, look to it not from a supplier - buyer horizontal relationship but rather a 360 degree angle where on any place around the circle all are business partners -whether clients, suppliers, distributors - all are like business partners (no capital is involved) and it has to be a win-win situation all the time for this to work". 242

262 Although the win-win situation makes these relationships with overseas ties more like business relationships, a high degree of trust and closeness is crucial (characteristics of dense networks). As will be explained in the following chapter Egyptian SMEs tend to have fewer overseas contacts and a denser network than UK SMEs who rely more on weak ties and larger networks over time. The percentage of foreign sales is positively correlated with number of regions in which the firm operates. In addition, the number of regions is highly correlated with the number of markets. A closer look at the nature of markets and regions in which Egyptian SMEs internationalise sheds some light on the reason behind this correlation. Some Egyptian SMEs choose to export to nearby counties especially in the Arab -Middle East region. This is due to the fact that there are similarities in culture, language and consumer tastes and preferences which makes it easier to enter into these markets and less reluctant to export to more distant markets. By contrast, those SMEs that export to a large number of diverse markets are more likely to higher percentage of foreign sales. Surprisingly, the number of years being international has no significant correlation with any of internationalisation outcomes or the network variables over time. This might imply that internationalisation experience of Egyptian SMEs is either of limited importance since they tend to internationalise to low psychic distance markets. Or it could be the main reason why they have limited internationalisation breadth compared to UK-based SMEs. Moreover, the yearly growth rate is positively correlated with average network density supporting the fact that network density plays a more crucial role in Egyptian SMEs internationalisation than does the network size. 243

263 As for the yearly international rate of growth the Egyptian SMEs has an average of 4% rate of growth and the average number of years being international is 14 years. Although it is the same yearly growth rate as the British SMEs, the number of years being international is much less than the UK sample with an average of 24 years. This lowers the rate of growth of British SMEs. So far, I have discussed the general evolutionary and co-evolutionary patterns of Egyptian and UK SMEs, without differentiating between the new venture and traditional SMEs. As discussed in Chapter 3, I proposed two a priori co-evolutionary frameworks for new ventures and traditional SMEs. The difference between these two types of firms is the speed of internationalisation the significance of which is explored in the following section. Speed of internationalisation as an alternative perspective of internationalisation process Firstly, in Chapter 2, I noted that 3 broad categories of SMEs were distinguished in the literature namely, traditional SMEs, born globals and born again globals (Bell, McNaughton, Young and Crick, 2003; Bell, McNaughton and Young, 2001; Kuivalainen, Saarenketo and Puumalainen, 2012). These categories have often been taken to designate different types of SMEs. However, they could be equally applied to indicate the difference between SMEs with respect to timing and speed of internationalisation because the main focus of internationalisation has been on the initiation or early stages of internationalisation. Few studies discuss what happens in subsequent stages. This means that question of the different patterns, in particular speed and timing of internationalisation, after its initiation has been scarcely addressed. 244

264 In two special issues of International Marketing Review this question has been addressed and different internationalisation patterns or pathways were identified in an attempt to better understand how SMEs internationalise and the different patterns companies pursue over time. However, the main focus of these special issues was again only on the first few years of SME internationalisation. An interesting exception was Kuivalainen et al., (2012) studied the internationalisation patterns of 78 Finnish SMEs and have collected data in 1999 from these SMEs and made a follow up round of data collection depending on secondary sources to identify what happened to these SMEs in The time period they covered was six years and they have identified several patterns of internationalisation for three types of SMEs, namely, BGs, BAGs and traditional SMEs. They focused on three main internationalisation process dimensions, mainly the scale, scope and timing of internationalisation 47. Following a similar logic, this present research attempts to investigate briefly how the different patterns of internationalisation Egyptian and UK SMEs follows with respect to the speed of internationalisation. In essence, the objective is to investigate if the integrating speed of internationalisation will offer an alternative explanation to the general patterns of internationalisation discussed in the above sections. I have added the time dimension in the previous discussion, through investigating over time the different internationalisation patterns that SMEs followed in terms of internationalisation scale, scope and geographical focus (breadth) (Yeoh, 2004; Jones, 2009) over their life cycle. However, investigating different internationalisation patterns of SMEs based on the speed of 47 The criteria for internationalisation scale they focused on was, BGs achieving at least 25% as foreign sales within the first three years of foundation). The scope of internationalisation they used to differentiate between Born Global companies and traditional SMEs is that BGs operate in at least five countries and finally they used the period of three years from SMEs inception as a cut-off period to differentiate between BGs and traditional SMEs (Knight and Cavusgil, 2004; Kuivalainen et al., 2012). 245

265 internationalisation and the subsequent internationalisation speed are suggested to offer additional insights. For the sake of simplicity, I have grouped the SMEs studied based on the timing of early internationalisation using the three years from company s inception -as a cut-off pointproposed by Knight and Cavusgil (2004) and Kuivalainen et al. (2012). For calculating the speed of subsequent internationalisation, I used an arbitrary cut-off point of six years. This is suggested to offer an indication of the likelihood of SME s survival (Oviatt and McDougall, 1994). However, I did not pre-determine a certain cut-off score for internationalisation scale and/or scope (For example, the 25% foreign sales criteria proposed by Knight and Cauvsgil, 2004). I wanted to identify the different internationalisation patterns based only on the speed of early and subsequent internationalisation. SMEs can vary in terms of their internationalisation scale and scope significantly across different points in time. Therefore, I sought to investigate whether those SMEs that undertake their first internationalisation activity soon after their inception, continued to internationalise rapidly in the subsequent internationalisation stage. Also, the question arises if SMEs start their initial internationalisation within the three years cut off point will they obtain more foreign sales revenues compared to traditional SMEs will they generate at least 25% of their sales revenues from overseas markets? How many markets and regions do they operate in? What happens in subsequent internationalisation stage will rapidly internationalising SMEs achieve higher percentage of foreign sales and operate in a large diverse number of markets compared to the traditional SMEs. Table 5.14 explains the classification of SMEs at different stages of internationalisation used in this section. 246

266 Table 5.14: Details of classification of SMEs at different stages of internationalisation SME type Early Start-up Traditional start-up Early internationalisation stage Subsequent internationalisation stage These are SMEs that started their first international activities within the first three years of the company's inception Early subsequent internationaliser These are SMEs that started their subsequent internationalisation stage within six years from company's inception These are SMEs that started their first international activities after three years of the company's inception Traditional subsequent internationaliser These are SMEs that started their subsequent internationalisation stage after six years from company's inception This classification is based only on timing or speed of early and subsequent internationalisation. I used the term "start up" to refer to first internationalisation attempts(s) and subsequent internationaliser to refer to the subsequent internationalisation attempts. Table 5.15 provides key descriptive of the UK and Egyptian SMEs based on the speed on internationalisation 48. Table 5.15: SMEs descriptives Type UK Egypt Early start-up Traditional start-up 8 7 Early subsequent internationaliser 7 10 Traditional subsequent internationaliser 11 3 None It important to note that some of the SMEs in both samples have not reached the SIS; this is why a third group (referred to as none) was created. 247

267 From the above Table 5.15, it clear that there are approximately the same number of early and traditional start up SMEs in both the UK and Egyptian samples. However, the number of SMEs that did not reach SIS is larger among Egyptian SMEs compared to those of the UK. The number of early subsequent internationalisers is higher among the Egyptian SMEs and the number of traditional subsequent internationalisers in higher among the UK SMEs. This raises an important question as to whether this means that the Egyptian SMEs are more likely to internationalise rapidly during SIS? In order to answer this question, it is important to look at other internationalisation outcomes or variables such as depth and breadth of internationalisation. Table 5.16 presents mean value of internationalisation outcomes of the early and traditional SMEs per stage. Table 5.16: Internationalisation outcomes for different types of SMEs Type of SME Internationalisation outcome UK Egypt Early start-up Percentage of foreign sales No. of markets No. of regions Traditional start-up Percentage of foreign sales Early subsequent internationaliser Traditional subsequent internationaliser No. of markets No. of regions Percentage of foreign sales No. of markets No. of regions Percentage of foreign sales No. of markets No. of regions

268 The findings presented in Table 5.16 shows that UK SMEs in general are more international across the three internationalisation dimensions compared to the Egyptian SMEs. Surprisingly, the there was a small variation between early subsequent internationalisers in the UK compared to Egypt in terms of PFS and the number of regions 49. Those early start up SMEs in both EIS and SIS achieve a high percentage of foreign sales (above the 25% criteria set by Knight and Cavusgil, 2004) and operate in a relatively large number of markets and in at least three regions. This confirms the literature that early starts up firms tend to seek from inception to internationalise into a large number of diverse overseas markets. However, as seen in Table 5.16 these early start-up SMEs do not necessarily obtain a higher percentage of foreign sales compared to traditional start-ups (for example, in the UK the average PFS is higher for the traditional start-ups) nor operate in more diverse markets s(for example, early and traditional start-ups in the UK operate nearly on average into the same number of regions). Moving to the SIS, UK early subsequent internationalisers have a higher percentage of foreign sales [PFS] than traditional subsequent internationalisers. However the latter operate in higher number of markets and regions. In contrast, Egyptian early subsequent internationalisers obtain a higher PFS and operate in a larger number of diverse markets compared to the traditional subsequent internationalisers. 49 I have taken the average of the highest percentage of foreign sales reported by SMEs for the first 6 years and after the first six years of firm s inception. This is to control for the age of the SME. 249

269 This raises an important question, do all early start-ups continue to internationalise rapidly in the SIS or do they slow down and tend to follow more of a traditional approach? In order to address this question a similar reasoning used by Kuivalainen et al, (2012) will be used to address the number of SMEs that continued to follow the same pattern-in terms of speed of internationalisation- they started with in EIS and/or if they changed their pattern in SIS. This is presented in Figures 5.6 and 5.7. Time >3 years <3 years Early start up Traditional start up (12) (8) Early Subsequent internationaliser (6) Traditional subsequent internationaliser (5) None (1) *(n) represents the number of companies Early Subsequent internationaliser (1) Traditional subsequent internationaliser (6) None (1) Source: Logic adopted from Kuivalainen et al.,(2012), but created by the author Figure 5.6: UK SMEs internationalisation patterns based on speed of early and subsequent internationalisation 250

270 Time >3 years <3 years Early start up Traditional start up (13) (7) Early Subsequent internationaliser (9) Traditional subsequent internationaliser (0) None (4) Early Subsequent internationaliser (1) Traditional subsequent internationaliser (3) None (3) Source: Logic adopted from Kuivalainen et al.,(2012), but created by the author Figure 5.7: Egyptian SMEs internationalisation patterns based on speed of early and subsequent internationalisation 251

271 Figures 5.6 and 5.7, show the different patterns emerging from the data. The early Start-up SMEs could follow any of the three following patterns into the SIS: 1) To continue to internationalise rapidly into further overseas markets. 2) To follow a rather slow internationalisation path (meaning that there is a long time lapse or gap between the timing of EIS and SIS). 3) Not entering the SIS, in other words, the SME is content with its current status quo. On the other hand, the Traditional start-up might follow any of the three following patterns: 1) It might accelerate its internationalisation after entering its first overseas markets. 2) It might continue to follow a slow internationalisation pattern (denoting a long time period between EIS and SIS) 3) It might not enter into the SIS. In other words, where the SME maintains its status quo. This shows that a company could start off its international operation within three years of its inception and later slow down its internationalisation. This could be the case in which an SME would follow a large client into several worldwide locations and only expand further after it has gained confidence and acquired experiential knowledge. The company might also maintain its status quo and not expand further into new markets because it generates a high percentage of foreign sales revenue from its existing overseas markets. Finally, SMEs might rapidly internationalise due to the help of its founder s dense and close relationships and has no further opportunities to expand further due to lack of new relationships to help them internationalise, or they might not be subjected to new sources of information the so-called closure effect (suggested by Burt, 1992, Colman 1988) and/or these companies especially 252

272 Egyptian companies might lack sufficient support from supporting agencies that would aid the company in its subsequent internationalisation. Furthermore, traditional start-ups or slow internationalisers might suddenly increase their level of internationalisation soon after they acquire needed experiential knowledge from both domestic and overseas operations. Moreover, some companies might continue to internationalise slowly suggesting incremental pattern of growth and finally these companies might not even enter SIS and maintain their status quo either because of lack of support, focus on few large clients and/or closure effect and limited access to structural holes. This is more evident in the case of Egyptian SMEs because the total number of SMES that did not enter into SIS is seven compared to only two cases in the UK). Studies on early start up or new ventures argued that Born Global companies operate in a large number of markets and achieve a higher percentage of foreign sales compared to traditional SMEs. To briefly assess the link between the speed of internationalisation (early and subsequent) Spearman correlations was carried out between timing of EIS and SIS and the three internationalisation outcomes in the EIS and the SIS. These are presented in Tables 5.17 and Table 5.17: Correlation between speed of internationalisation and internationalisation outcomes of UK SMEs (Spearman's rho correlation coefficient) Time EIS Time SIS Time EIS Time SIS Percentage of foreign sales t 2 Percentage foreign sales t 3 Number of markets t 2 Number of markets t 3 Number of regions t 2 Number of regions t * * * * *Correlation is significant at the 0.05 level (2-tailed). **Correlation is significant at the 0.02 level (2-tailed). *** N=20 253

273 Table 5.18: Correlation between speed of internationalisation and internationalisation outcomes of Egyptian SMEs (Spearman's rho correlation coefficient) Time EIS Time SIS Percentage of foreign sales t 2 Percentage of foreign sales t 3 Number of markets t 2 Number of markets t 3 Number of regions t 2 Number of regions t 3 Time EIS Time SIS ** ** * * *Correlation is significant at the 0.05 level (2-tailed). **Correlation is significant at the 0.02 level (2-tailed). *** N=20 From Table 5.17 shows that the timing of EIS of UK SMEs is positively correlated with timing of SIS. This means that if an SME starts to internationalise soon after its inception it is more likely to rapidly internationalise in the subsequent internationalisation stage. Also if the SME internationalises slowly it is more likely to move into SIS after a long period of time. As shown in Figure 5.6, nearly half of the UK early start ups and traditional start-ups continued to follow the same pattern as they did in EIS. However, Table 5.19 shows that among the Egyptian SMEs there is no significant correlation between the timing of internationalisation in the EIS and the SIS. In addition, the timing of SIS of UK and Egyptian SMEs is positively correlated with at least two of the internationalisation outcomes in SIS, which suggests that the longer it takes SME to enter SIS, the more likely it will operate in a larger number of markets and regions. In the case of Egyptian SMEs it is more likely to achieve higher PFS in SIS and vice versa. 254

274 Finally, the number of markets in EIS is negatively correlated with timing of SIS for the Egyptian SMEs. This implies that the larger number of markets the SME operates in EIS, the longer it will take the company to enter into SIS. The next question is whether there is a correlation between speed of internationalisation and network variables in in the EIS and the SIS. Tables 5.19 and 5.20 present the correlations between speed of internationalisation and network variables. Table 5.19: Correlation between speed of internationalisation and network variables of UK SMEs (Spearman's rho correlation coefficient) Time EIS Time SIS Network density t 2 Network density Network size Net work size Network constraint Network constraint t 3 t 2 t 3 t 2 t 3 Time EIS Time SIS * * * * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.02 level (2-tailed). *** N=20 Table 5.20: Correlation between speed of internationalisation and network variables of Egyptian SMEs (Spearman's rho correlation coefficient) Time EIS Time SIS Network density t 2 Network density t 3 Network size t 2 Network size t 3 Network constraint t 2 Network constraint t 3 Time EIS Time SIS ** ** ** *Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.02 level (2-tailed). *** N=20 255

275 Table 5.19 shows that the timing of SIS is positively correlated with network size in SIS for the UK SMEs. This suggests that the longer it takes the company to reach the SIS the more likely it will have more contacts in its network at that stage. This suggests that companies that are slow to subsequently internationalise are more likely to have larger networks. On the other hand, Table 5.20 shows the timing of the EIS is negatively correlated with network size at EIS for Egyptian SMEs. This implies that the more rapid the SME s early internationalisation the more likely it will have smaller networks, perhaps focusing more on dense and pre-existing relationships of its founder. The timing of SIS is positively correlated with network density and network constraint in the SIS. This suggests that more rapidly the SME is entering into SIS, the more likely it will have more sparse networks characterised by a large number of bridging relationships and the slower the company is to enter into SIS the more likely it will have more dense networks in that stage. This means that companies tend to rely more on bridging relationships in order to gain faster access to more markets hence increasing their internationalisation or they rely on a dense network which might create a closure effect and slow down the speed of the SME s subsequent internationalisation.alternatively, companies tend to be more satisfied with their status quo generating enough sales revenue for either domestic or few overseas markets (few large clients). Finally it could be the lack of support that makes companies to firstly assess whether their personal social network would offer means and support needed to internationalise and if not they would then tend to create and enter new markets either via creating new relationships or through a bridging relationship. 256

276 Discussion and conclusions In this chapter, I have identified the general patterns of association between network variables and internationalisation outcomes, within and across stages. The purpose was to identify general tendencies of how different network variables correlated one another, how internationalisation outcomes correlated with each another and finally how network and internationalisation variables/outcomes co-influence one another. The co-evolutionary patterns were identified within stages and across stages. This provided a general overview on which variables co-evolve together. These findings have generated the following patterns for UK and Egyptian SMEs as shown in Figures It important to note that the opposite of the correlations presented in figures 5.7 and 5.8 is also true. For example, if two variables are positively correlated this indicates that if one increases the other will increase and if one decreases the other will decrease. 257

277 Co-evolutionary Model of UK SMEs Pre-internationalisation stage Early internationalisation stage Subsequent internationalisation stage Internationalisation Process No. of markets ( H) No. of regions (H) No. of markets ( H) No. of regions (H) Networking Process Network constraint (L) i..e Bridging relationships (H) Network Size (H) Network Size ( H4) Over time Legend: Av. Network Size (H) No. of markets Tn (H) No. Of regions Tn (H) (H) (L) Co-evolution Line High Low Figure 5.8: UK SMEs co-evolutionary Model 258

278 Co-evolutionary Model of Egyptian SMEs Pre-internationalisation stage Early internationalisation stage Subsequent internationalisation stage Internationalisation Process PFS (H) No. of markets ( H) No. of regions (H) PFS (H) No. of regions (H) No. of Markets (H) -ve Networking Process Net Density (H) Net Constraint (H) i,e bridging relationships (L) Network Size (H) Network Density (L) Network Constraint (L) i.e Bridging relationships (H) Network Size ( H) Network Density(H) Network Constraint (H) Over time Av. Network Density (H) Av. Network Constraint (H) I.e Bridging relationships (L) Figure 5.9: Egyptian SMEs co-evolution model No. of markets Tn (H) PFS Tn (H) No. Of regions Tn (H) Legend: Co-evolution Line (H) High (L) Low 259

279 As shown in Figure 5.8, there is a positive correlation between average network size and the number of markets and regions in which UK SMEs operate. This is a co-evolutionary pattern across stages which offer an overview of how the average network size over the three stages is associated with total number of markets and regions in which the SME operates at time tn. This involved the accumulation of number of markets in which the company operates during EIS and SIS 51. This shows that among the three network variables only network size was associated significantly with internationalisation spread and scope across the stages. Taking each stage individually, network size is positively correlated with number of markets and regions in which SME operates during the EIS and SIS. Moreover, network constraint during PIS is negatively correlated with number of markets and regions in SIS. This means that companies rely on bridging relationships they had during the PIS to increase their internationalisation spread and scope in subsequent internationalisation stage. The Egyptian SMEs exhibited a different co-evolutionary pattern as shown in figure 5.9. The average network density and constraint were positively correlated with the three internationalisation outcomes at time tn, showing the importance of dense and close networks in the internationalisation process of Egyptian SMEs. Similar patterns could be identified within stages. For example, network density and constraint are positively correlated with PFS, number of regions and markets in which SMEs operates in SIS. Moreover, the network density and limited number of bridging relationships SMEs had during the PIS played a significant role it in SMEs subsequent internationalisation. This is shown by the positive 51 To avoid duplication, the number of markets through which company de-internationalized is not counted except during the particular stage in which company entered into those markets. For instance, if a given SME entered a particular market in the EIS and further penetrated this market in the SIS this particular market is counted once. 260

280 correlation between the network density and constraint level in the PIS and the three internationalisation outcomes in SIS. Surprisingly, during the EIS, network size was the most important network variable as it was positively correlated with the internationalisation outcomes. Furthermore, network constraint and density were both negatively correlated with number of regions and markets in the SIS. This suggests that if SMEs have less dense networks composed of bridging relationship in EIS, they are more likely to operate in a larger number of diverse markets in the SIS and vice versa. This also implies that if they have dense networks in the EIS they will operate in less number of markets and regions in the SIS. It is important to consider the opposite scenario of any proposed pattern of association as these are based on correlations result so that for instance, a positive relationship between two variables this simply indicates that they increase or decrease together. It is therefore important to understand the underlying processes behind the correlations using on more indepth data. In the following chapter, I explored in more detail the underlying mechanisms and dynamic nature of the SMEs networking and internationalisation processes and how they co-influence and complement one another. i The rationale behind sampling and criteria of sampling were presented at the 2012 BAM conference paper. This paper was about the evolutionary patterns of NVs authored by researcher and supervisor (Narooz and Child, 2012). Moreover, some of the methodological aspects mentioned in this chapter have been reported in one of the assignments submitted by the researcher as part of her first year methodology module in June

281 CHAPTER 6 SMEs TAXONOMIES AND CLUSTERS Introduction In the last chapter, I examined the general evolutionary and co-evolutionary patterns of UK and Egyptian SMEs. This focused on the patterns of association between different network variables and internationalisation performance/outcome variables. The chapter showed some general tendencies and linkages between different variables rather than explaining the underlying mechanisms and processes. It emerged that in the case of Egyptian SMEs, there was a significant positive correlation between the degree of network cohesiveness (measured by the network density) and the percentage of foreign sales to total sales. By contrast, the UK SMEs demonstrated a significant positive correlation between the size of network and the number of overseas markets in which UK SMEs operate. This shows a general tendency of UK entrepreneurs to focus more on operating in a large number of markets and relying on a large number of network ties. By contrast, the Egyptian SMEs tend to focus more on cohesive (dense) networks in order to generate more sales from a limited number of overseas markets. These general tendencies provide us with overview of the internationalisation and networking behaviours of UK and Egyptian SMEs. The aim of this chapter is to understand the underlying processes and mechanisms of SME internationalisation and networking behaviour over time. To this end, I carried out an exploratory clustering of the UK and Egyptian SMEs in an attempt to pinpoint general 262

282 similarities and differences between individual SMEs within each national sample. 52 This clustering is based on average scores across stages of internationalisation. I subsequently extend the analysis to take account the changes over time. The clustering of SMEs is based on two main variables; namely the average network size (Av. Net size) and the number of overseas markets in which a given SME operates. The rationale behind choosing these two variables is twofold. Firstly, based on the results of the Mann Whitney U test results, which showed that there is a significant difference between the UK and Egyptian SMEs networking and internationalisation behaviour in terms of average network size and the number of overseas markets in which the firm operates. Secondly, I use average network size and number of markets rather than network density and percentage of foreign sales because average network size offers a better understanding of the extent to which ego (an SME) invests in building network relationships that are crucial for its international growth. However, average network density refers to the extent to which ego s alters are connected to one another, which if taken separately from network size could lead to some confusion. For example, two SMEs might have a similar network density score of 30%; while one of them has a network size of 50 and the other has a network size of just 10. Both have 30% of their network ties connected to one another (which means three ties of the latter SME are connected to one another and 15 ties of the former are connected to one another). It is clear that the first SME (which has 50 ties) invests more in building relationships and creating new ties than the second SME. Furthermore, SMEs could derive the bulk of their percentage of foreign sales from one market; which is a major limitation of using percentage of foreign sales as the main indicator 52 The sampling possibilities that were available particularly in Egypt did not permit for a systematic comparison of other potential contingences such as sector differences and whether the SMEs are a traditional, knowledge intensive or knowledge based. The limited number of Egyptian SMEs that operate overseas and the scarcity of knowledge based SMEs reduced the scope for systematic comparison between two contexts. 263

283 of the SMEs degree of internationalisation 53. However, using the number of markets in which company operates indicates the extent to which the company is scope and extensively involved in internationalisation. This chapter will start by discussing the results of Mann Whitney U test regarding the key differences between two samples in terms of across stage co-evolutionary measures; I then perform exploratory clustering of UK and Egyptian SMEs using the hierarchical clustering technique. This generates six different clusters three from each sample. The resulting taxonomies for each national sample will be discussed in detail. Generic networking and internationalisation differences between UK and Egypt SMEs As a starting point, Mann Whitney U tests were used to assess whether there are significant differences between UK and Egyptian SMEs networking and internationalisation behaviours. A non-parametric equivalent to T-test (Mann Whitney U test) is used to compare two independent samples based on different across stage network and internationalisation variables. The Mann-Whitney U test was carried out between the two samples to test if there are significant differences between networking and internationalisation variables or not 54. The level of significance used here is "Monte-Carlo exact test" rather than the asymptotic method, 53 As shown in the last chapter, the number of markets is correlated with the number of regions in which a company operates. Moreover, within each cluster, I have looked at concentration in one region versus diversification to different regions as another indicator of SMEs degree of international scope. 54 As expected given the small sample size in both the UK and Egypt cases, the test of homogeneity of variance was carried out for each sample individually and there was no significant variance within each sample for the normality test to be computed. As mentioned in the previous chapter, tests for normality such as the Shapiro-Wilk test were used for both samples. Among the variables used in chapter 5 five variables approximated normal distribution these were mainly average network density and average network size in both UK and Egypt and the number of regions in the UK sample) approximated normal distribution. The remaining seven variables were not normally disturbed. This explains the need to use a non-parametric test to see if there is a significant difference between the two samples. Therefore, I used Mann-Whitney U to test for significant differences between the UK and Egypt (Field, 2011). The results of the Shapiro-Wilk test and the test of homogeneity of variance (Levene s test) are in Appendix

284 because the exact test is recommended for small sample size (Field, 2011:547). These results are shown in Tables 1 and 2 in Appendix The average network density of the UK SMEs (Median=8.05) 55 did not differ significantly from that of the Egyptian SMEs (Median=6.70), U=197.5, Z=-0.068,ns, r= Moreover, the average network constraint of UK SMEs (Median=0.29) did not differ significantly from Egyptian SMEs (Median=0.23), U=137, z=-1.704,ns, r= However, the average network size of the UK SMEs (Median=17.5) tended to be significantly larger than that of the Egyptian SMEs (Median= 13), U=111, z=-2.409, p<0.05, r= As for the internationalisation outcomes, the percentage of foreign sales (at time Tn) of the UK SMEs (Median = 55%) did not differ significantly from that of Egyptian SMEs (Median =50%), U=183.5, z=-.448, ns, r= Also, the number of regions (at time Tn) in which UK SMEs operate (Median =5.00) did not differ significantly from that of Egyptian SMEs (Median = 4.00), U=152.5, Z=-1.309, ns, r= However, the number of overseas markets (at time Tn) in which UK SMEs operate (Mdn= 22) differed significantly from Egyptian SMEs (Mdn=10), U=79.5, Z=-3.270, p<0.05, r=-0. The UK SMEs tended to be active in a larger number of overseas markets. 55 In these results, the medians are of the distribution of individual SME scores averaged over the three stages of internationalisation. 56 The effect size donated by r is computed to give a standardized measure of the size of the effect observed in this research so that it could be compared in other studies(field, 2011). It follows this formula: r=z/ N, where z=z-score and N is the sample size. 265

285 Accordingly, the UK SMEs differed significantly from the Egyptian SEMs in terms of the number of contacts inherent in the SMEs networks and the number of overseas markets in which they operate. However, UK SMEs did not differ significantly in terms of the percentage of foreign sales and the average network density and constraint. One plausible explanation is the way in which the network density and constraint are calculated as a proportion of the network size. Hence they depend on the network size and one could argue that it is the network size that makes the difference because the density and constraint are contingent to some extent on the number of alters that are present in ego s network. Moreover, the majority of foreign revenues (percentage of foreign sales) could be obtained from one overseas market or from several. Again, one could argue that the number of overseas markets in which the SMEs operate gives a clearer indication of the scope of the SMEs internationalisation. One of the key pitfalls of the quantitative measures is that they do not uncover the underlying mechanisms and dynamics of network and internationalisation behaviours of the UK and Egyptian SMEs (Coveillo, 2005; Bryman and Bell, 2003). In order to gain a clearer and more in depth insights into the networking and internationalisation behaviour of UK and Egyptian SMEs, a more detailed analysis of the two samples is warranted. As shown in Chapter 5, UK SMEs exhibited different networking behaviours over time compared to the Egyptian SMEs over time. More specifically, they tend to focus more on creating and increasing their network size rather than on nurturing and maintaining existing strong relationships (i.e. they focus to a lesser extent on maintaining strong cohesive networks). Moreover, UK SMEs tend to focus more on enhancing the breath of internationalisation (in terms of the number of overseas markets and the number of regions) than Egyptian SMEs. This was evident in the across stages co-evolutionary correlation between the network size and number of overseas markets in which UK SMEs operate. 266

286 On the other hand, Egyptian SMEs focused more on nurturing and cultivating their dense networks and on enhancing the depth of internationalisation than did the UK SMEs. This was also evident in the across stages co-evolutionary correlations between the percentage of foreign sales and the network density. These tendencies are put forward based on identifying the key significant correlations between different network and internationalisation variables. However, is there a significant difference between the two samples? A good starting point is to see whether there is a significant difference between the UK and Egyptian SMEs, in terms of their networking behaviour (in terms of network density, constraint and size) and internationalisation focus (percentage of foreign sales, number of markets and number of regions). As just reported, the Mann Whitney U test was carried out based on these variables to pinpoint if there is a significant difference between the two samples or not. Consequently, I undertook a deeper investigation of the dynamics of the co-evolution of the networking and internationalisation processes over time. For the sake of simplicity and to make the discussion more manageable, a preliminary exploratory clustering was carried out on each sample. Common characteristics of the companies within each sample are highlighted and exemplary cases are identified and discussed. A. Proposed SMEs clustering In an attempt to explore the data and to come up with empirically derived groups of SMEs, an exploratory clustering technique is used to arrive at taxonomies of Egyptian and UK SMEs respectively. Since this research is based on a small number of SMEs in each sample and an unknown number of clusters, agglomerative hierarchical clustering is used (Hair, Black, Babin, Anderson and Tatham, 2005; Mooi and Sarstedt, 2011). 267

287 The agglomerative hierarchical clustering technique is a commonly used technique especially for small sample sizes (Mooi and Sarstedt, 2011; Hair et al., 2005). The clustering algorism-which is a measure of (dis)similarity is based on the squared Euclidean distance and the clustering method used is complete linkage algorism which is a useful technique if the researcher wants to depict outliers and to have a small compact clusters (Mooi and Sarstedt, 2011; Hair et al., 2005) 57. Hierarchical clustering generates "n-1" clustering solutions (when n is equal to the number of cases). It involves combining or adding up cases into a "treelike structure" (Hair et al., 2005:584). There are two types of hierarchical clustering, namely agglomerative and divisive clustering. The former involves building up a tree or hierarchy whereby each case in the first stage is grouped into one cluster and the next stage (dis)similar cases are grouped together into one cluster and so on until all clusters are grouped into one main cluster. The latter involves having all cases grouped into one main cluster and then in each subsequent step these cases are divided into clusters, until each case is represented in a cluster on its own (Hair el al., 2005). The average network size and number of markets of UK SMEs differed significantly from those of Egyptian SMEs across the two internationalisation stages. It is therefore appropriate to investigate whether the clustering SMEs based on these two variables would enhance our understanding of the underlying mechanisms and dynamics of SMEs networking and internationalisation behaviours. However, it is important to take into account the issue of multi-collinearity or high correlation between the clustering variables (i.e correlation coefficient R> 0.09) as this 57 The choice of Euclidean distance as a measure of the degree of similarity or dissimilarity between objects is one of the most commonly used measures. Moreover, the UK sample has one or two outliers. I decided not to omit or remove the outliers but rather sought to understand the reasons behind their unique behaviour. 268

288 might make these variables over represented in the clusters (Mooi and Sarstedt, 2011; Hair et al., 2005). To check for the multi-collinearity problem, the spearman correlation is used to test if the two variables have a correlation coefficient of more than The correlation coefficients of the average network size and the number of markets of both the UK and Egyptian SMEs were below The correlation matrix is shown in Appendix 2.3. As a general guide, the number of clustering variables depends on a sample size of 2 n, whereby n=number of clustering variables (Formann, 1984; Mooi and Sarstedt, 2011). For example, 2 4 =16, which is the minimum number of cases to be present if 4 clustering variables were to be used in clustering cases. In this study the only two clustering variables will be used for the two samples of (20 SMEs each) which is roughly a reasonable sample. Hierarchical clustering was carried out in two stages. In the first stage, the technique was used without specifying any clustering solution to come up with a dendrogram and an agglomerative matrix. The dendrogram and agglomerative matrix provide a tentative guide showing all the possible clustering solutions for each sample. Moreover, the decision regarding which clustering solution to use depended on theoretical and practical reasons. In this study the objective has been to come up with tentative clusters of SMEs that shows varying degrees of networking and internationalisation behaviours and having a manageable number of clusters that is sufficient to show within group differences and similarities is warranted. The second stage, involved re-running the analysis while specifying the number of clusters to be used. Accordingly cluster membership is provided and each case is then placed in its corresponding cluster across the two dimensions chosen as clustering variables. 269

289 The next section will illustrate the clustering of UK cases and Egyptian cases separately. A.1 Clustering of UK SMEs The UK SMEs were clustered based on the average network size and number of overseas markets at time tn. The first step of the analysis involved identifying the possible number of clustering solutions and deciding on the best number of clusters to be used. 58 This is done through examining dendrogram presented in Figure 6.1 which is an SPSS generated dendrogram with possible clustering solutions of UK SMEs 59. Case Figure 6.1: UK SMEs dendrogram. 58 Given the differences in the measures and scales of network size and number of markets, the values for each variable were standardized using z-score. 59 Agglomeration matrix for both UK and Egyptian clustering solutions are in appendix

290 According to Figure 6.1, three or four cluster solutions emerge. Moreover, the number of clusters is also contingent upon logical and theoretical reasoning (Hair et al, 2005). A three cluster solution seems appropriate based on practical and theoretical considerations. Having three clusters showing varying degrees of internationalisation and networking extensity - ranging from high, moderate and low- would serve the purpose of this exploratory technique and would help the researcher to identify key within-group similarities and between group differences. After re-running the analysis with the three cluster solution the following cluster of UK SMEs were identified based on the two clustering variables (namely average network size and number of overseas markets at tn). These clusters are shown in figure

291 SM Manu1 No. of Overseas Markets (Internationa lisation Spread) Limited Internationaliser/ networker Lek Manu12 Manu8 Manu6 Manu3 Manu10 Con2 SCE Manu7 LW1 Manu5 Manu4 Manu2 Manu9 OG1 7-9 LC Con1 OG2 Extensive Internationaliser/networker Moderate Average Network Size (Network Extensitivity) Figure 6.2: UK SMEs taxonomy A.2 Clustering of Egyptian SMEs The Egyptian SMEs were also clustered based on their average network size and the number of overseas markets at time tn. The first step of the analysis involved identifying the possible number of clustering solutions and deciding on the best number of clusters to be used 60. Examining the dendrogram provides guidelines as to the number of clusters to be used. 60 Given the differences in the measures and scales of network size and number of overseas markets, the values for each variable were standardized using z-score. 272

292 These are presented in Figure 6.3; an SPSS generated dendrogram with possible clustering solutions of Egyptian SMEs. Figure 6.3: Egyptian SMEs dendrogram. In a similar fashion to the UK cases, a closer look at figure 6.3 indicated that three or four cluster solutions were possible. For sake of comparability, the three cluster solution will be used. 273